ORIGINAL ARTICLE
Histologic Categorization of Fibrotic Cancer Stroma in the Primary Tumor Is an Independent Prognostic Index in Resectable Colorectal Liver Metastasis Hideki Ueno, MD, PhD,* Tsuyoshi Konishi, MD,w Yuichi Ishikawa, MD,z Hideyuki Shimazaki, MD,y Masashi Ueno, MD,w Suefumi Aosasa, MD,* Akio Saiura, MD,w Kazuo Hase, MD,* and Junji Yamamoto, MD*
Abstract: Although the molecular mechanism of desmoplastic reaction (DR) for providing aggressive tumor characteristics is increasingly recognized, the prognostic role of DR has not been investigated in colorectal liver metastasis (CRLM). A pathologic review of 412 patients who underwent hepatectomy for CRLM at 2 independent institutions was conducted. DR in primary tumors was classified as mature, intermediate, or immature on the basis of the existence of keloid-like collagen and myxoid stroma–distinctive histologic products of extracellular matrix remodeling. With respect to DR, 137, 122, and 153 patients were classified as mature, intermediate, and immature, respectively. Immature DRs were associated with higher T and N stages, higher primary tumor grade, synchronous and larger size of liver metastasis, and extrahepatic disease (Pr0.0001 to 0.002). DR significantly influenced the rate of recurrence in extrahepatic sites, including the lung, peritoneum, and local region in the primary tumor (Pr0.0001 to 0.03), rather than the remnant liver. Five-year overall survival rates after hepatectomy were the highest in the mature group (58.9%), followed by intermediate (42.1%) and immature (26.7%) groups. A significant prognostic impact of DR was observed in subset analyses for institutions, primary tumor location, and timing and number of liver metastases. Multivariate analysis revealed that DR was an independent prognostic factor along with T stage of the primary tumor, size of liver metastasis, and extrahepatic disease. Characterizing DR in the primary tumor on the basis of histologic products of cancer-associated fibroblasts is valuable in evaluating prognostic outcome after hepatectomy in CRLM patients. Key Words: colorectal liver metastasis, desmoplastic reaction, cancer-associated fibroblasts, extracellular matrix, cancer microenvironment (Am J Surg Pathol 2014;38:1380–1386) From the Departments of *Surgery; yLaboratory Medicine, National Defense Medical College, Saitama; wDepartment of Gastroenterological Surgery; and zDivision of Pathology, Cancer Institute Hospital, Tokyo, Japan. Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article. Correspondence: Hideki Ueno, MD, PhD, Department of Surgery, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
1380 | www.ajsp.com
P
atients with resectable colorectal liver metastases (CRLM) represent a heterogenous population of oncological outcome; some patients experience recurrent disease and die within a short period of time after hepatectomy, whereas some live without recurrence of disease. This is because of the great diversity in the potential of growth and metastasis in individual tumors as well as in the anatomic extent of the disease in this group of patients. In terms of the anatomic extent of the disease, quantitative parameters such as the number1–5 and size of liver metastases,4–6 extrahepatic disease,5 primary tumor nodal status,2–5,7 and serum carcinoembryonic antigen levels1–4 have been repeatedly reported for their relevance to prognoses after hepatectomy. Some of these have been used to develop risk stratification systems for the resection of CRLM8,9 despite criticism that their broad application has limited value with respect to patient stratification for clinical management in controversial areas, such as the administration of chemotherapy or surveillance programs.10 In contrast, parameters for grading the potential of primary tumor aggressiveness have not been sufficiently investigated and have rarely been included in proposed prognostic scoring systems.11 The major reason for this is probably the identification of few valuable histoprognostic parameters comparable to parameters for the anatomic extent of the disease, and primary tumor parameters are likely to be regarded as having an insignificant prognostic value in cases of metastatic colorectal cancer. For examples, tumor differentiation grade and vascular invasion in the primary tumor, which are the representative indexes determining postoperative treatment management in curatively resected colorectal cancer cases,12 have no reproducible prognostic value in resectable CRLM cases.2,5 Recent basic studies in tumor biology have shown that the tumor microenvironment plays a major role in modulating the metastatic capacity of most cancers and determines the oncological outcome of tumors.13,14 Fibroblasts and myofibroblasts, the most abundant mesenchymal cells within most carcinomas, promote tumor progression and are present in the histopathologic Am J Surg Pathol
Volume 38, Number 10, October 2014
Am J Surg Pathol
Volume 38, Number 10, October 2014
entity termed the desmoplastic reaction (DR).15,16 Although DR in the primary tumor has not been the focus of investigation in CRLM, characterization of the individual tumor microenvironment is expected to provide important prognostic information after hepatectomy, independent of the anatomic extent of the disease, because it is now recognized that tumor stromal cells are the essential contributors to the growth, survival, invasiveness, and metastatic ability of neoplastic epithelial cells within these tumors.15 We have previously reported that categorizing DR in the primary tumor on the basis of distinctive histologic features of the products of activated fibroblasts, namely keloid-like collagen and myxoid stroma, provided prognostic information in colorectal cancer patients.17,18 In the present study, we used this method to histologically characterize the tumor microenvironment and investigate its value in predicting the prognostic outcome in patients with curatively resected CRLM at 2 individual institutions.
PATIENTS AND METHODS Patients A total of 412 colorectal cancer patients who underwent potentially curative surgery for their primary tumors and synchronous (n = 222) and metachronous (n = 190) liver metastases at 2 independent institutions, the National Defense Medical College Hospital (1984 to 2010) and the Cancer Institute Hospital (1995 to 2007),
Primary Tumor Desmoplasia in CRLM
were included. The average age was 61.4 years (range, 28 to 91 y), and 272 (66.0%) patients were men. With an average follow-up period of 61 months (range, 11 to 188 mo) for survivors, the overall survival (OS) and disease-free survival rates at 5 years after hepatectomy were 41.9% and 21.8%, respectively. A retrospective review of the medical records including operative, pathology, and follow-up reports was conducted. In addition, one of the authors (H.U.) pathologically reviewed the primary tumors to evaluate the pattern of DR with no knowledge of patients’ clinical outcomes.
Histologic Categorization of DR Hematoxylin and eosin–stained glass slides prepared from a single longitudinal section of the whole tumor, including its deepest part, were microscopically scanned to identify the area including keloid-like collagen or myxoid stroma. Keloid-like collagens were broad bundles of hypocellular collagen with bright eosinophilic hyalinization, typically observed in keloid.17 Myxoid stroma is defined as an amorphous stromal substance composed of an amphophilic or slightly basophilic extracellular matrix (ECM) substance, which is usually intermingled with randomly oriented keloid-like collagen.17 These components are mostly observed in the reactive fibrous zone at the advancing edge of a tumor. DR was histologically classified as 1 of 3 categories (mature, intermediate, and immature) on the basis of the existence of keloid-like collagen or myxoid stroma (Fig. 1).18 More specifically, it was regarded as mature
FIGURE 1. Categorization of DR in the primary tumor. The pattern of DR in the reactive fibrous zone at the leading edge of the primary tumor is classified in 3 categories according to the presence or absence of keloid-like collagen or myxoid stroma. A, Mature type: fibrotic stroma characterized by fine fibers stratified into multilayers without keloid-like collagen and myxoid stroma. B, Intermediate type: fibrotic stroma characterized by keloid-like collagen (broad bands of collagen with bright eosinophilic hyalinization, similar to those observed in a keloid) intermingled with mature fibers. C, Immature type: abundant ECM substance accumulated at the leading edge of the tumor to form myxoid stroma (hematoxylin and eosin staining). r
2014 Lippincott Williams & Wilkins
www.ajsp.com |
1381
Ueno et al
Am J Surg Pathol
Volume 38, Number 10, October 2014
TABLE 1. Categorization of DR in the Primary Tumor and Other Prognostic Factors in Patients Undergoing Hepatectomy for CRLM No. Patients (%) According to DR Category Prognostic Factors Primary tumor Tumor differentiation T N Lymphatic invasion Venous invasion Liver metastasis Timing of metastasis Diameter Number Extrahepatic disease CEA*
Categories
Mature
Intermediate
Immature
P < 0.0001
G1 G2 G3 T1,2 T3 T4 Negative Positive Absence Presence Absence Presence
83 49 5 21 98 18 60 77 29 108 37 100
(40.5) (25.8) (29.4) (100) (34.1) (17.3) (50.0) (26.4) (53.7) (30.2) (27.0) (28.9)
69 (33.7) 50 (26.3) 3 (17.7) 0 96 (33.4) 26 (25.0) 33 (27.5) 89 (30.5) 16 (29.6) 106 (29.6) 17 (25.8) 105 (30.3)
53 (25.9) 91 (47.9) 3 (52.9) 0 93 (32.4) 60 (57.7) 27 (22.5) 126 (43.2) 9 (16.7) 144 (40.2) 12 (18.2) 141 (40.8)
Synchronous Metachronous r5 cm > 5 cm 1-4 Z5 Absence Presence r5.0 mg/L > 5.0 mg/L
62 75 100 37 120 17 134 3 42 83
(27.9) (39.5) (30.1) (46.3) (35.0) (24.6) (35.2) (9.7) (32.8) (33.1)
61 61 105 17 100 22 116 6 42 74
99 54 127 26 123 30 131 22 44 94
(27.5) (32.1) (31.6) (21.3) (29.2) (31.9) (30.4) (19.4) (32.8) (29.5)
(44.6) (28.4) (38.3) (32.5) (35.9) (43.5) (34.4) (71.0) (34.4) (37.5)
< 0.0001 < 0.0001 0.0006 < 0.0001 0.0024 0.0187 0.2355 0.0002 0.7676
*Among 379 patients for whom the preoperative CEA level was known. CEA indicates carcinoembryonic antigen at hepatectomy.
when the fibrotic stroma did not contain keloid-like collagen or myxoid stroma and was only composed of fine mature collagen fibers stratified into multiple layers. When keloid-like collagens were intermingled in mature stroma, typically having parallel orientation to the mature collagen fibers, fibrotic stroma was defined as intermediate mature. Fibrotic stroma with myxoid changes was regarded as immature. The stromal assessment in each case was classified according to the most immature stromal area. Stroma around microscopic abscesses was excluded from consideration.
Statistical Analyses Survival rates were calculated using the KaplanMeier method, and comparisons were made using the log rank test. After categorization, each clinical and pathologic variable was entered into the Cox proportional hazard regression analysis to determine which parameter had an independent effect on the postoperative survival. The associations among prognostic factors and their associations with recurrence were analyzed by w2 tests. Statistical analyses were performed with SPSS Statistics 17.0 (SPSS Inc., Chicago, IL) and StatView ver. 5.0 (SAS Institute Inc., Cary, NC).
RESULTS Incidence of Each DR Category DR was classified as mature, intermediate, or immature for 137, 122, and 153 tumors, respectively. DR
1382 | www.ajsp.com
was significantly associated with various clinicopathologic prognostic factors—that is, immature stroma was relevant to unfavorable categories in each prognostic factor, including T and N stages, primary tumor histology (tumor grade, lymphatic and venous invasion), the timing and diameter of liver metastasis, and extrahepatic disease (Pr0.0001 to 0.002; Table 1).
Prognostic Impact of DR Categorization According to the DR category, 2-and 5-year disease-free survival rates after hepatectomy were 46.4% and 35.2% for mature stroma, 39.2% and 24.9% for intermediate stroma, and 19.0% and 7.5% for immature stroma, respectively (Fig. 2, P < 0.0001). Similarly, 5year OS rates after hepatectomy were the highest in the mature group (58.9%), followed by intermediate (42.1%) and immature groups (26.7%). Table 2 shows subset analyses for the prognostic value of DR categorization. The significant impact of DR categorization on OS was similar in both institutions. In both the institutions, patients in the mature DR group had the most favorable survival results, and the presence of immature stroma had the most unfavorable impact on survival, followed by the presence of intermediate stroma. Similarly, the significant impact of DR categorization on OS was observed in subset analyses of the timing of metastasis, primary tumor location, and number of liver metastases. r
2014 Lippincott Williams & Wilkins
Am J Surg Pathol
Volume 38, Number 10, October 2014
DFS
Recurrence Pattern According to DR Categorization
OS
Probability
Probability
1.0
1.0
0.5
0.5
Primary Tumor Desmoplasia in CRLM
Mature (137) Mature (137) Intermediate (122)
DR categorization had a significant correlation with the incidence of recurrence in extrahepatic organs (P < 0.0001) but had no significant correlation with the incidence of recurrence in the remnant liver (Table 3). With regard to the recurrence pattern, the incidence of recurrence in the lung, peritoneum, and local site significantly increased with a decrease in the maturation of fibrotic stroma (Pr0.0001 to 0.03).
Intermediate (122) Immature (153) P< 0.0001
0 0
Immature (153)
1 2 3 4 5 Years after hepatectomy
0
Multivariate Analyses for OS
P< 0.0001 0
1 2 3 4 5 6 7 8 Years after hepatectomy
FIGURE 2. Survival estimates of patients undergoing hepatectomy for colorectal liver metastases using the KaplanMeier method according to the categorization of DR in the primary tumor. DFS: mature versus intermediate, P = 0.1440; mature versus immature, P < 0.0001; intermediate versus immature, P = 0.0003. OS: mature versus intermediate, P = 0.0027; mature versus immature, P < 0.0001; intermediate versus immature, P = 0.0084. DFS indicates disease-free survival. Numbers in parentheses are the number of patients.
Among potential prognostic factors, the parameters that significantly impacted OS after hepatectomy were as follows: T and N stages in the primary tumor, tumor differentiation, lymphatic invasion, DR, the diameter of liver metastasis, and extrahepatic disease (Table 4). The timing of liver metastasis, number of liver metastases, and serum carcinoembryonic antigen level at hepatectomy had a marginal significance. Among these, multivariate analysis based on the Cox proportional hazard model identified DR categorization as an independent prognostic marker of OS in addition to
TABLE 2. Subset Analyses for the Impact of DR in the Primary Tumor on OS After Hepatectomy for CRLM Characteristics Institution
Subset Institution 1 Institution 2
Timing of metastasis
Synchronous Metachronous
Primary tumor location
Colon Rectum
Depth of invasion
T3 T4
Number of liver metastasis
Single Multiple
DR Category (No. Cases) Mature (61) Intermediate (53) Immature (92) Mature (76) Intermediate (69) Immature (61) Mature (62) Intermediate (61) Immature (99) Mature (75) Intermediate (61) Immature (54) Mature (91) Intermediate (73) Immature (82) Mature (46) Intermediate (49) Immature (71) Mature (98) Intermediate (96) Immature (93) Mature (18) Intermediate (26) Immature (60) Mature (75) Intermediate (63) Immature (59) Mature (62) Intermediate (59) Immature (94)
Cox Proportional Hazard Model 5-y Survival (%) 59.2 36.5 21.2 58.4 48.0 36.5 58.5 40.8 29.5 58.2 44.3 23.5 54.4 36.5 32.8 69.4 48.7 19.6 57.6 46.5 31.9 44.2 26.7 18.2 61.2 43.8 17.5 56.0 40.7 32.9
P
HR (95% CI) 1.6 2.7 1.8 2.2 2.1 2.6 1.4 2.5 1.8 2.2 1.7 3.2 1.5 2.2 1.8 2.0 1.6 3.0 1.9 2.4
1 (1.0-2.7) (1.7-4.3) 1 (1.1-3.0) (1.3-3.7) 1 (1.3-3.6) (1.6-4.3) 1 (0.9-2.4) (1.5-4.1) 1 (1.1-2.8) (1.4-3.4) 1 (0.9-3.2) (1.8-5.6) 1 (1.0-2.3) (1.4-3.3) 1 (0.8-4.0) (1.0-4.2) 1 (1.0-2.7) (1.9-4.9) 1 (1.1-3.2) (1.4-3.9)
0.0666 < 0.0001 0.0226 0.0037 0.0054 < 0.0001 0.1774 < 0.0001 0.0122 < 0.0001 0.0887 < 0.0001 0.0700 < 0.0001 0.1649 0.0521 0.0676 < 0.0001 0.0182 < 0.0001
CI indicates confidence interval; HR, hazard ratio.
r
2014 Lippincott Williams & Wilkins
www.ajsp.com |
1383
Ueno et al
Am J Surg Pathol
Volume 38, Number 10, October 2014
TABLE 3. Impact of DR Categorization in the Primary Tumor on the Recurrence After Hepatectomy for CRLM No. Patients With Recurrence (%) According to the DR Category Recurrence (overall) Patterns of recurrence Remnant liver Extrahepatic sites Lung Lymph node Peritoneal dissemination Local (primary site) Bone/brain
Mature
Intermediate
Immature
P
88 (64.2)
88 (72.1)
137 (89.5)
< 0.0001
61 44 28 9 6 5 7
51 57 30 15 6 12 5
81 103 67 24 18 19 11
0.1477 < 0.0001 < 0.0001 0.0522 0.0263 0.0269 0.5175
(44.5) (32.1) (20.4) (6.6) (4.4) (3.7) (5.1)
extrahepatic disease, the diameter of liver metastasis, and the T stage of the primary tumor (Table 4).
DISCUSSION Oncological research for cancer metastasis has traditionally focused on the properties of malignant cells. However, recent studies in tumor biology have shown that the tumor microenvironment exerts a major influence on tumor behavior, including the metastatic process.13,14 Among the different cell types that constitute the
(41.8) (46.7) (24.6) (12.3) (4.9) (9.8) (4.1)
(52.9) (67.3) (43.8) (15.7) (11.8) (12.4) (7.2)
tumor microenvironment, “cancer-associated fibroblasts” (CAFs), the term subsuming at least 2 distinct cell types (cells with similarities to the fibroblasts that create the structural foundation supporting most normal epithelial tissues and myofibroblasts),16 have attracted increasing attention both as recipients and producers of protumorigenic signals. Several lines of evidence demonstrate that CAFs essentially contribute to the microenvironmental regulation of tumor metastasis.13–15,19 DR, which is a histologic response of fibrotic cancer stroma modulated by CAFs, is a hallmark of invasive colorectal
TABLE 4. Univariate and Multivariate Analyses for OS in Patients Undergoing Hepatectomy for CRLM Univariate Analysis Prognostic Factors Primary tumor T N Tumor differentiation Lymphatic invasion Venous invasion DR Liver metastasis Timing of metastasis Diameter Number
Multivariate Analysis
Categories (No. Cases)
5-y Survival (%)
P
HR (95% CI)
T1/T2 (21) T3 (287) T4 (104) Negative (120) Positive (292) G1 (17) G2 (205) G3 (190) Negative (54) Positive (358) Negative (66) Positive (346) Mature (137) Intermediate (122) Immature (153)
80.0 45.6 24.5 54.7 36.6 48.6 35.5 34.0 60.1 39.2 43.0 41.6 58.9 42.1 26.7
< 0.0001 0.0006
1 2.5 (0.8-8.1) 4.0 (1.2-13.1) NS
0.0012
NS
—
0.0312
NS
—
0.8327
NS
—
1 1.6 (1.1-2.3) 2.0 (1.4-2.9)
0.0192 < 0.0001
Synchronous (222) Metachronous (190) < 5 cm (332) Z5 cm (80) < 5 (343) Z5 (69)
40.6 43.5 43.5 35.0 43.1 35.3
0.0912
NS
—
0.0250
1 1.5 (1.1-2.0) NS
0.0213 —
Absence (381) Presence (31)
44.7 7.2
< 0.0001
1 2.0 (1.3-3.0)
0.0018
r5.0 mg/L (128) > 5.0 mg/L (251)
48.1 39.5
0.0573
< 0.0001
0.0518
P
0.1236 0.0232 —
Extrahepatic disease CEA*
NS
—
*Among 379 patients for whom the preoperative CEA level was known. CEA indicates carcinoembryonic antigen at hepatectomy; CI, confidence interval; HR, hazard ratio; NS, not selected.
1384 | www.ajsp.com
r
2014 Lippincott Williams & Wilkins
Am J Surg Pathol
Volume 38, Number 10, October 2014
cancer15,20; however, it is morphologically diverse according to individual tumors, possibly reflecting the wide spectrum of CAF functions. Previously, we proposed a simple 3-tiered histologic categorization of DR in the reactive fibrous zone on the basis of a histologic review and prognostic analyses of rectal cancer cases treated at St Marks Hospital.17,18 As yardsticks of criteria for the system, 2 distinctive histologic features of DR were adopted, that is, keloid-like collagen and myxoid stroma, both of which can be diagnosed on hematoxylin and eosin–stained slides. Principally, both stromal components do not appear in noninvasive tumors and are observed at the leading edge of invasive colorectal cancer wherein dynamic interaction between tumor cells and host cells, mediated by cytokines, chemokines, or growth factors, is embodied as a form of ECM remodeling. Keloid-like collagen is a term used for thick bands of hyalinized collagen in the reactive fibrous zone of colorectal cancer, which are morphologically similar to those in keloids–thick scars of the human skin or cornea, produced by the deposition of excessive amounts of collagen over prolonged periods.21 Pathophysiologically, a keloid is characterized by fibroblast proliferation and the synthesis of ECM components through the overexpression of growth factors, such as transforming growth factor b.22,23 The distinctive appearance of keloidlike collagens, which are immunohistochemically characterized by an excessive deposition of type 1 collagen,17 indicates the deviant activation of CAFs in the local tumor microenvironment. Myxoid stroma is formed as a result of extensive ECM remodeling and is characterized by a massive deposition of nonfibrillar ECM. Stromal cells, including myofibroblasts and bone marrow–derived cells, play a crucial role in ECM remodeling as sources of various types of protease activity, including matrix metalloproteases, cathepsins, and plasminogen activators.13,24 It is increasingly appreciated that ECM remodeling by such protease activity facilitates tumor invasion and metastasis by cleaving cell-adhesion molecules, by paving a path through the degradation of ECM, and by activating growth factors and cytokines.13,19 Myxoid stroma is also characterized by inconspicuous lymphocyte infiltration,17,18 indicating that the area is in a state of immunosuppression and may favor tumor progression. The St Marks Hospital series analyzed 862 curatively resected rectal cancer cases and found that the 5year survival rate was highest in the mature group (80%), followed by the intermediate (55%) and immature groups (27%).18 In the present study, we observed consistent results in both the independent institutions in terms of the adverse prognostic impact of the unfavorable DR pattern in the primary tumor: the rate of survival after hepatectomy was the highest in the mature group and lowest in the immature group. The significant prognostic impact was similarly observed in subset analyses for the primary tumor location, number of liver metastases, and, notably, liver metastasis timing. Multivariate analyses in the r
2014 Lippincott Williams & Wilkins
Primary Tumor Desmoplasia in CRLM
combined data set revealed that DR categorization was associated with prognostic outcome, independent of conventional prognostic parameters reflecting the anatomic extent of the disease in CRLM patients, such as the number and size of metastatic tumors and extrahepatic disease. It is becoming evident that characterizing the tumor microenvironment can provide prognostic information14; however, this is the first study to demonstrate that characterizing the microenvironment on the basis of the histologic component of ECMs, produced by CAFs as the result of the tumor-host interface, can predict oncological outcome in CRLM cases. Our study demonstrated that DR categorization was associated with the incidence of extrahepatic recurrence, particularly in the lung, peritoneum, and local region in the primary tumor, rather than recurrence in the remnant liver. The significance of pursuing prognostic markers in the primary tumors in cases of metastatic colorectal cancers is largely undetermined. However, the proportion of patients with recurrence limited to the remnant liver is only 30% to 40% in patients with recurrence after liver resection,25 and patient prognosis dominantly depends on the occurrence of extrahepatic metastasis originating from the primary tumor. It is therefore reasonable to assume that there are prognostic determinants in primary tumors, even in the CRLM setting. The molecular mechanisms and clinical advantages of histologic evaluation of the primary tumor microenvironment in CRLM cases remain to be elucidated in detail. Reportedly, DR categorization is significantly associated with the mismatch-repair status26; however, details on the genetic background of morphologic alterations of DR in colorectal cancers remain to be clarified. Therefore, this histologic categorization method can be used as a simple methodology to understand the status of the cancer microenvironment. Furthermore, recent studies have indicated that the microenvironment can affect the sensitivity of malignant cells to chemotherapy: tumor cells became less sensitive to chemotherapy following coculture with CAFs27,28 or CAF-derived ECM proteins.29 In addition, the accumulation of ECM components conceivably compresses blood and lymphatic vessels, leading to reduced perfusion and ultimately impeding the delivery of drugs to neoplastic cells.30 Further studies should focus on the correlation between histologic DR categorization and drug sensitivity and on whether metastatic tumors form DR patterns similar to those in the primary site. In conclusion, we found that the histologic characterization of the primary tumor microenvironment on the basis of products of CAFs helps in evaluating prognostic outcomes after hepatectomy in CRLM patients. The oncological benefit of neoadjuvant chemotherapy in patients suitable for curative hepatectomy remains debatable.31–33 The information obtained by comprehensively evaluating the microenvironmental regulation of metastases demonstrated in the primary tumor and the anatomical extent of the disease may allow us to determine the risk of www.ajsp.com |
1385
Ueno et al
Am J Surg Pathol
relapse after hepatectomy and aid in optimal patient selection for such a controversial treatment. REFERENCES 1. Sasaki A, Iwashita Y, Shibata K, et al. Analysis of preoperative prognostic factors for long-term survival after hepatic resection of liver metastasis of colorectal carcinoma. J Gastrointest Surg. 2005;9:374–380. 2. Minagawa M, Yamamoto J, Kosuge T, et al. Simplified staging system for predicting the prognosis of patients with resectable liver metastasis. Arch Surg. 2007;142:269–276. 3. Lee W-S, Kim MJ, Yun SH, et al. Risk factor stratification after simultaneous liver and colorectal resection for synchronous colorectal metastasis. Langenbecks Arch Surg. 2008;393:13–19. 4. Merkel S, Bialecki D, Meyer T, et al. Comparison of clinical risk scores predicting prognosis after resection of colorectal liver metastases. J Surg Oncol. 2009;100:349–357. 5. Beppu T, Sakamoto Y, Hasegawa K, et al. A nomogram predicting disease-free survival in patients with colorectal liver metastases treated with hepatic resection: multicenter data collection as a Project Study for Hepatic Surgery of the Japanese Society of Hepato-Biliary-Pancreatic Surgery. J Hepatobiliary Pancreat Sci. 2012;19:72–84. 6. Tanaka K, Shimada H, Ueda M, et al. Long-term characteristics of 5-year survivors after liver resection for colorectal metastases. Ann Surg Oncol. 2007;14:1336–1346. 7. Tan MCB, Castaldo ET, Gao F, et al. A prognostic system applicable to patients with resectable liver metastasis from colorectal carcinoma staged by positron emission tomography with [18F]fluoro2-deoxy-D-glucose: role of primary tumor variables. J Am Coll Surg. 2008;206:857–869. 8. Nordlinger B, Guiguet M, Vaillant J-C, et al. Surgical resection of colorectal carcinoma metastases to the liver: a prognostic scoring system to improve case selection, based on 1568 patients. Cancer. 1996;77:1254–1262. 9. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases. Ann Surg. 1999;230:309–321. 10. Gomez D, Cameron IC. Prognostic scores for colorectal liver metastasis: clinically important or an academic exercise? HPB (Oxford). 2010;12:227–238. 11. Ueno H, Mochizuki H, Hatsuse K, et al. Indicators for treatment strategies of colorectal liver metastases. Ann Surg. 2000;231:59–66. 12. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology-colon cancer (version 3. 2014) 2012. Available at: http://www.nccn.org/professionals/physician_gls/pdf/ colon.pdf. Accessed February 1, 2014. 13. Joyce JA, Pollard JW. Microenvironmental regulation of metastasis. Nat Rev Cancer. 2009;9:239–252. 14. Allen M, Jones L. Jekyll and Hyde: the role of the microenvironment on the progression of cancer. J Pathol. 2011;223:162–176. 15. McAllister SS, Weinberg RA. Tumor-host interactions: a farreaching relationship. J Clin Oncol. 2010;28:4022–4028.
1386 | www.ajsp.com
Volume 38, Number 10, October 2014
16. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–674. 17. Ueno H, Jones A, Jass JR, et al. Clinicopathological significance of the ‘keloid-like’ collagen and myxoid stroma in advanced rectal cancer. Histopathology. 2002;40:327–334. 18. Ueno H, Jones AM, Wilkinson KH, et al. Histological categorisation of fibrotic cancer stroma in advanced rectal cancer. Gut. 2004;53:581–586. 19. Pietras K, O¨stman A. Hallmarks of cancer: interactions with the tumor stroma. Exp Cell Res. 2010;316:1324–1331. 20. Hewitt RE, Powe DG, Carter GI, et al. Desmoplasia and its relevance to colorectal tumour invasion. Int J Cancer. 1993;53: 62–69. 21. O’Sullivan ST, O’Connor TPF, O’Shaughnessy M. Aetiology and management of hypertrophic scars and keloids. Ann R Coll Surg Engl. 1996;78:168–175. 22. Bran GM, Goessler UR, Hormann K, et al. Keloids: current concepts of pathogenesis (review). Int J Mol Med. 2009;24:283–293. 23. Robles DT, Moore E, Draznin M, et al. Keloids: pathophysiology and management. Dermatol Online J. 2007;13:9. 24. Desmoulie`re A, Guyot C, Gabbiani G. The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol. 2004;48:509–517. 25. Ueno H, Mochizuki H, Hashiguchi Y, et al. Predictors of extrahepatic recurrence after resection of colorectal liver metastases. Br J Surg. 2004;53:581–586. 26. Wright C, Stewart I. Histopathology and mismatch repair status of 458 consecutive colorectal carcinomas. Am J Surg Pathol. 2003;27:1393–1406. 27. Mu¨erko¨ster S, Wegehenkel K, Arlt A, et al. Tumor stroma interactions induce chemoresistance in pancreatic ductal carcinoma cells involving increased secretion and paracrine effects of nitric oxide and interleukin-1beta. Cancer Res. 2004;64:1331–1337. 28. Shekhar MPV, Santner S, Carolin KA, et al. Direct involvement of breast tumor fibroblasts in the modulation of tamoxifen sensitivity. Am J Pathol. 2007;170:1546–1560. 29. Ahmed AA, Mills AD, Ibrahim AEK, et al. The extracellular matrix protein TGFBI induces microtubule stabilization and sensitizes ovarian cancers to paclitaxel. Cancer Cell. 2007;12:514–527. 30. Feig C, Gopinathan A, Neesse A, et al. The pancreas cancer microenvironment. Clin Cancer Res. 2012;18:4266–4276. 31. Nordlinger B, Sorbye H, Glimelius B, et al. Perioperative chemotherapy with FOLFOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): a randomised controlled trial. Lancet. 2008;371: 1007–1016. 32. Adam R, Bhangui P, Poston G, et al. Is perioperative chemotherapy useful for solitary, metachronous, colorectal liver metastases? Ann Surg. 2010;252:774–787. 33. Chua TC, Saxena A, Liauw W, et al. Systematic review of randomized and nonrandomized trials of the clinical response and outcomes of neoadjuvant systematic chemotherapy for resectable colorectal liver metastases. Ann Surg Oncol. 2010;17:492–501.
r
2014 Lippincott Williams & Wilkins
Histologic Categorization of Fibrotic Cancer Stroma in the Primary Tumor Is an Independent Prognostic Index in Resectable Colorectal Liver Metastasis Hideki Ueno, MD, PhD,* Tsuyoshi Konishi, MD,w Yuichi Ishikawa, MD,z Hideyuki Shimazaki, MD,y Masashi Ueno, MD,w Suefumi Aosasa, MD,* Akio Saiura, MD,w Kazuo Hase, MD,* and Junji Yamamoto, MD*
Abstract: Although the molecular mechanism of desmoplastic reaction (DR) for providing aggressive tumor characteristics is increasingly recognized, the prognostic role of DR has not been investigated in colorectal liver metastasis (CRLM). A pathologic review of 412 patients who underwent hepatectomy for CRLM at 2 independent institutions was conducted. DR in primary tumors was classified as mature, intermediate, or immature on the basis of the existence of keloid-like collagen and myxoid stroma–distinctive histologic products of extracellular matrix remodeling. With respect to DR, 137, 122, and 153 patients were classified as mature, intermediate, and immature, respectively. Immature DRs were associated with higher T and N stages, higher primary tumor grade, synchronous and larger size of liver metastasis, and extrahepatic disease (Pr0.0001 to 0.002). DR significantly influenced the rate of recurrence in extrahepatic sites, including the lung, peritoneum, and local region in the primary tumor (Pr0.0001 to 0.03), rather than the remnant liver. Five-year overall survival rates after hepatectomy were the highest in the mature group (58.9%), followed by intermediate (42.1%) and immature (26.7%) groups. A significant prognostic impact of DR was observed in subset analyses for institutions, primary tumor location, and timing and number of liver metastases. Multivariate analysis revealed that DR was an independent prognostic factor along with T stage of the primary tumor, size of liver metastasis, and extrahepatic disease. Characterizing DR in the primary tumor on the basis of histologic products of cancer-associated fibroblasts is valuable in evaluating prognostic outcome after hepatectomy in CRLM patients. Key Words: colorectal liver metastasis, desmoplastic reaction, cancer-associated fibroblasts, extracellular matrix, cancer microenvironment (Am J Surg Pathol 2014;38:1380–1386) From the Departments of *Surgery; yLaboratory Medicine, National Defense Medical College, Saitama; wDepartment of Gastroenterological Surgery; and zDivision of Pathology, Cancer Institute Hospital, Tokyo, Japan. Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article. Correspondence: Hideki Ueno, MD, PhD, Department of Surgery, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
1380 | www.ajsp.com
P
atients with resectable colorectal liver metastases (CRLM) represent a heterogenous population of oncological outcome; some patients experience recurrent disease and die within a short period of time after hepatectomy, whereas some live without recurrence of disease. This is because of the great diversity in the potential of growth and metastasis in individual tumors as well as in the anatomic extent of the disease in this group of patients. In terms of the anatomic extent of the disease, quantitative parameters such as the number1–5 and size of liver metastases,4–6 extrahepatic disease,5 primary tumor nodal status,2–5,7 and serum carcinoembryonic antigen levels1–4 have been repeatedly reported for their relevance to prognoses after hepatectomy. Some of these have been used to develop risk stratification systems for the resection of CRLM8,9 despite criticism that their broad application has limited value with respect to patient stratification for clinical management in controversial areas, such as the administration of chemotherapy or surveillance programs.10 In contrast, parameters for grading the potential of primary tumor aggressiveness have not been sufficiently investigated and have rarely been included in proposed prognostic scoring systems.11 The major reason for this is probably the identification of few valuable histoprognostic parameters comparable to parameters for the anatomic extent of the disease, and primary tumor parameters are likely to be regarded as having an insignificant prognostic value in cases of metastatic colorectal cancer. For examples, tumor differentiation grade and vascular invasion in the primary tumor, which are the representative indexes determining postoperative treatment management in curatively resected colorectal cancer cases,12 have no reproducible prognostic value in resectable CRLM cases.2,5 Recent basic studies in tumor biology have shown that the tumor microenvironment plays a major role in modulating the metastatic capacity of most cancers and determines the oncological outcome of tumors.13,14 Fibroblasts and myofibroblasts, the most abundant mesenchymal cells within most carcinomas, promote tumor progression and are present in the histopathologic Am J Surg Pathol
Volume 38, Number 10, October 2014
Am J Surg Pathol
Volume 38, Number 10, October 2014
entity termed the desmoplastic reaction (DR).15,16 Although DR in the primary tumor has not been the focus of investigation in CRLM, characterization of the individual tumor microenvironment is expected to provide important prognostic information after hepatectomy, independent of the anatomic extent of the disease, because it is now recognized that tumor stromal cells are the essential contributors to the growth, survival, invasiveness, and metastatic ability of neoplastic epithelial cells within these tumors.15 We have previously reported that categorizing DR in the primary tumor on the basis of distinctive histologic features of the products of activated fibroblasts, namely keloid-like collagen and myxoid stroma, provided prognostic information in colorectal cancer patients.17,18 In the present study, we used this method to histologically characterize the tumor microenvironment and investigate its value in predicting the prognostic outcome in patients with curatively resected CRLM at 2 individual institutions.
PATIENTS AND METHODS Patients A total of 412 colorectal cancer patients who underwent potentially curative surgery for their primary tumors and synchronous (n = 222) and metachronous (n = 190) liver metastases at 2 independent institutions, the National Defense Medical College Hospital (1984 to 2010) and the Cancer Institute Hospital (1995 to 2007),
Primary Tumor Desmoplasia in CRLM
were included. The average age was 61.4 years (range, 28 to 91 y), and 272 (66.0%) patients were men. With an average follow-up period of 61 months (range, 11 to 188 mo) for survivors, the overall survival (OS) and disease-free survival rates at 5 years after hepatectomy were 41.9% and 21.8%, respectively. A retrospective review of the medical records including operative, pathology, and follow-up reports was conducted. In addition, one of the authors (H.U.) pathologically reviewed the primary tumors to evaluate the pattern of DR with no knowledge of patients’ clinical outcomes.
Histologic Categorization of DR Hematoxylin and eosin–stained glass slides prepared from a single longitudinal section of the whole tumor, including its deepest part, were microscopically scanned to identify the area including keloid-like collagen or myxoid stroma. Keloid-like collagens were broad bundles of hypocellular collagen with bright eosinophilic hyalinization, typically observed in keloid.17 Myxoid stroma is defined as an amorphous stromal substance composed of an amphophilic or slightly basophilic extracellular matrix (ECM) substance, which is usually intermingled with randomly oriented keloid-like collagen.17 These components are mostly observed in the reactive fibrous zone at the advancing edge of a tumor. DR was histologically classified as 1 of 3 categories (mature, intermediate, and immature) on the basis of the existence of keloid-like collagen or myxoid stroma (Fig. 1).18 More specifically, it was regarded as mature
FIGURE 1. Categorization of DR in the primary tumor. The pattern of DR in the reactive fibrous zone at the leading edge of the primary tumor is classified in 3 categories according to the presence or absence of keloid-like collagen or myxoid stroma. A, Mature type: fibrotic stroma characterized by fine fibers stratified into multilayers without keloid-like collagen and myxoid stroma. B, Intermediate type: fibrotic stroma characterized by keloid-like collagen (broad bands of collagen with bright eosinophilic hyalinization, similar to those observed in a keloid) intermingled with mature fibers. C, Immature type: abundant ECM substance accumulated at the leading edge of the tumor to form myxoid stroma (hematoxylin and eosin staining). r
2014 Lippincott Williams & Wilkins
www.ajsp.com |
1381
Ueno et al
Am J Surg Pathol
Volume 38, Number 10, October 2014
TABLE 1. Categorization of DR in the Primary Tumor and Other Prognostic Factors in Patients Undergoing Hepatectomy for CRLM No. Patients (%) According to DR Category Prognostic Factors Primary tumor Tumor differentiation T N Lymphatic invasion Venous invasion Liver metastasis Timing of metastasis Diameter Number Extrahepatic disease CEA*
Categories
Mature
Intermediate
Immature
P < 0.0001
G1 G2 G3 T1,2 T3 T4 Negative Positive Absence Presence Absence Presence
83 49 5 21 98 18 60 77 29 108 37 100
(40.5) (25.8) (29.4) (100) (34.1) (17.3) (50.0) (26.4) (53.7) (30.2) (27.0) (28.9)
69 (33.7) 50 (26.3) 3 (17.7) 0 96 (33.4) 26 (25.0) 33 (27.5) 89 (30.5) 16 (29.6) 106 (29.6) 17 (25.8) 105 (30.3)
53 (25.9) 91 (47.9) 3 (52.9) 0 93 (32.4) 60 (57.7) 27 (22.5) 126 (43.2) 9 (16.7) 144 (40.2) 12 (18.2) 141 (40.8)
Synchronous Metachronous r5 cm > 5 cm 1-4 Z5 Absence Presence r5.0 mg/L > 5.0 mg/L
62 75 100 37 120 17 134 3 42 83
(27.9) (39.5) (30.1) (46.3) (35.0) (24.6) (35.2) (9.7) (32.8) (33.1)
61 61 105 17 100 22 116 6 42 74
99 54 127 26 123 30 131 22 44 94
(27.5) (32.1) (31.6) (21.3) (29.2) (31.9) (30.4) (19.4) (32.8) (29.5)
(44.6) (28.4) (38.3) (32.5) (35.9) (43.5) (34.4) (71.0) (34.4) (37.5)
< 0.0001 < 0.0001 0.0006 < 0.0001 0.0024 0.0187 0.2355 0.0002 0.7676
*Among 379 patients for whom the preoperative CEA level was known. CEA indicates carcinoembryonic antigen at hepatectomy.
when the fibrotic stroma did not contain keloid-like collagen or myxoid stroma and was only composed of fine mature collagen fibers stratified into multiple layers. When keloid-like collagens were intermingled in mature stroma, typically having parallel orientation to the mature collagen fibers, fibrotic stroma was defined as intermediate mature. Fibrotic stroma with myxoid changes was regarded as immature. The stromal assessment in each case was classified according to the most immature stromal area. Stroma around microscopic abscesses was excluded from consideration.
Statistical Analyses Survival rates were calculated using the KaplanMeier method, and comparisons were made using the log rank test. After categorization, each clinical and pathologic variable was entered into the Cox proportional hazard regression analysis to determine which parameter had an independent effect on the postoperative survival. The associations among prognostic factors and their associations with recurrence were analyzed by w2 tests. Statistical analyses were performed with SPSS Statistics 17.0 (SPSS Inc., Chicago, IL) and StatView ver. 5.0 (SAS Institute Inc., Cary, NC).
RESULTS Incidence of Each DR Category DR was classified as mature, intermediate, or immature for 137, 122, and 153 tumors, respectively. DR
1382 | www.ajsp.com
was significantly associated with various clinicopathologic prognostic factors—that is, immature stroma was relevant to unfavorable categories in each prognostic factor, including T and N stages, primary tumor histology (tumor grade, lymphatic and venous invasion), the timing and diameter of liver metastasis, and extrahepatic disease (Pr0.0001 to 0.002; Table 1).
Prognostic Impact of DR Categorization According to the DR category, 2-and 5-year disease-free survival rates after hepatectomy were 46.4% and 35.2% for mature stroma, 39.2% and 24.9% for intermediate stroma, and 19.0% and 7.5% for immature stroma, respectively (Fig. 2, P < 0.0001). Similarly, 5year OS rates after hepatectomy were the highest in the mature group (58.9%), followed by intermediate (42.1%) and immature groups (26.7%). Table 2 shows subset analyses for the prognostic value of DR categorization. The significant impact of DR categorization on OS was similar in both institutions. In both the institutions, patients in the mature DR group had the most favorable survival results, and the presence of immature stroma had the most unfavorable impact on survival, followed by the presence of intermediate stroma. Similarly, the significant impact of DR categorization on OS was observed in subset analyses of the timing of metastasis, primary tumor location, and number of liver metastases. r
2014 Lippincott Williams & Wilkins
Am J Surg Pathol
Volume 38, Number 10, October 2014
DFS
Recurrence Pattern According to DR Categorization
OS
Probability
Probability
1.0
1.0
0.5
0.5
Primary Tumor Desmoplasia in CRLM
Mature (137) Mature (137) Intermediate (122)
DR categorization had a significant correlation with the incidence of recurrence in extrahepatic organs (P < 0.0001) but had no significant correlation with the incidence of recurrence in the remnant liver (Table 3). With regard to the recurrence pattern, the incidence of recurrence in the lung, peritoneum, and local site significantly increased with a decrease in the maturation of fibrotic stroma (Pr0.0001 to 0.03).
Intermediate (122) Immature (153) P< 0.0001
0 0
Immature (153)
1 2 3 4 5 Years after hepatectomy
0
Multivariate Analyses for OS
P< 0.0001 0
1 2 3 4 5 6 7 8 Years after hepatectomy
FIGURE 2. Survival estimates of patients undergoing hepatectomy for colorectal liver metastases using the KaplanMeier method according to the categorization of DR in the primary tumor. DFS: mature versus intermediate, P = 0.1440; mature versus immature, P < 0.0001; intermediate versus immature, P = 0.0003. OS: mature versus intermediate, P = 0.0027; mature versus immature, P < 0.0001; intermediate versus immature, P = 0.0084. DFS indicates disease-free survival. Numbers in parentheses are the number of patients.
Among potential prognostic factors, the parameters that significantly impacted OS after hepatectomy were as follows: T and N stages in the primary tumor, tumor differentiation, lymphatic invasion, DR, the diameter of liver metastasis, and extrahepatic disease (Table 4). The timing of liver metastasis, number of liver metastases, and serum carcinoembryonic antigen level at hepatectomy had a marginal significance. Among these, multivariate analysis based on the Cox proportional hazard model identified DR categorization as an independent prognostic marker of OS in addition to
TABLE 2. Subset Analyses for the Impact of DR in the Primary Tumor on OS After Hepatectomy for CRLM Characteristics Institution
Subset Institution 1 Institution 2
Timing of metastasis
Synchronous Metachronous
Primary tumor location
Colon Rectum
Depth of invasion
T3 T4
Number of liver metastasis
Single Multiple
DR Category (No. Cases) Mature (61) Intermediate (53) Immature (92) Mature (76) Intermediate (69) Immature (61) Mature (62) Intermediate (61) Immature (99) Mature (75) Intermediate (61) Immature (54) Mature (91) Intermediate (73) Immature (82) Mature (46) Intermediate (49) Immature (71) Mature (98) Intermediate (96) Immature (93) Mature (18) Intermediate (26) Immature (60) Mature (75) Intermediate (63) Immature (59) Mature (62) Intermediate (59) Immature (94)
Cox Proportional Hazard Model 5-y Survival (%) 59.2 36.5 21.2 58.4 48.0 36.5 58.5 40.8 29.5 58.2 44.3 23.5 54.4 36.5 32.8 69.4 48.7 19.6 57.6 46.5 31.9 44.2 26.7 18.2 61.2 43.8 17.5 56.0 40.7 32.9
P
HR (95% CI) 1.6 2.7 1.8 2.2 2.1 2.6 1.4 2.5 1.8 2.2 1.7 3.2 1.5 2.2 1.8 2.0 1.6 3.0 1.9 2.4
1 (1.0-2.7) (1.7-4.3) 1 (1.1-3.0) (1.3-3.7) 1 (1.3-3.6) (1.6-4.3) 1 (0.9-2.4) (1.5-4.1) 1 (1.1-2.8) (1.4-3.4) 1 (0.9-3.2) (1.8-5.6) 1 (1.0-2.3) (1.4-3.3) 1 (0.8-4.0) (1.0-4.2) 1 (1.0-2.7) (1.9-4.9) 1 (1.1-3.2) (1.4-3.9)
0.0666 < 0.0001 0.0226 0.0037 0.0054 < 0.0001 0.1774 < 0.0001 0.0122 < 0.0001 0.0887 < 0.0001 0.0700 < 0.0001 0.1649 0.0521 0.0676 < 0.0001 0.0182 < 0.0001
CI indicates confidence interval; HR, hazard ratio.
r
2014 Lippincott Williams & Wilkins
www.ajsp.com |
1383
Ueno et al
Am J Surg Pathol
Volume 38, Number 10, October 2014
TABLE 3. Impact of DR Categorization in the Primary Tumor on the Recurrence After Hepatectomy for CRLM No. Patients With Recurrence (%) According to the DR Category Recurrence (overall) Patterns of recurrence Remnant liver Extrahepatic sites Lung Lymph node Peritoneal dissemination Local (primary site) Bone/brain
Mature
Intermediate
Immature
P
88 (64.2)
88 (72.1)
137 (89.5)
< 0.0001
61 44 28 9 6 5 7
51 57 30 15 6 12 5
81 103 67 24 18 19 11
0.1477 < 0.0001 < 0.0001 0.0522 0.0263 0.0269 0.5175
(44.5) (32.1) (20.4) (6.6) (4.4) (3.7) (5.1)
extrahepatic disease, the diameter of liver metastasis, and the T stage of the primary tumor (Table 4).
DISCUSSION Oncological research for cancer metastasis has traditionally focused on the properties of malignant cells. However, recent studies in tumor biology have shown that the tumor microenvironment exerts a major influence on tumor behavior, including the metastatic process.13,14 Among the different cell types that constitute the
(41.8) (46.7) (24.6) (12.3) (4.9) (9.8) (4.1)
(52.9) (67.3) (43.8) (15.7) (11.8) (12.4) (7.2)
tumor microenvironment, “cancer-associated fibroblasts” (CAFs), the term subsuming at least 2 distinct cell types (cells with similarities to the fibroblasts that create the structural foundation supporting most normal epithelial tissues and myofibroblasts),16 have attracted increasing attention both as recipients and producers of protumorigenic signals. Several lines of evidence demonstrate that CAFs essentially contribute to the microenvironmental regulation of tumor metastasis.13–15,19 DR, which is a histologic response of fibrotic cancer stroma modulated by CAFs, is a hallmark of invasive colorectal
TABLE 4. Univariate and Multivariate Analyses for OS in Patients Undergoing Hepatectomy for CRLM Univariate Analysis Prognostic Factors Primary tumor T N Tumor differentiation Lymphatic invasion Venous invasion DR Liver metastasis Timing of metastasis Diameter Number
Multivariate Analysis
Categories (No. Cases)
5-y Survival (%)
P
HR (95% CI)
T1/T2 (21) T3 (287) T4 (104) Negative (120) Positive (292) G1 (17) G2 (205) G3 (190) Negative (54) Positive (358) Negative (66) Positive (346) Mature (137) Intermediate (122) Immature (153)
80.0 45.6 24.5 54.7 36.6 48.6 35.5 34.0 60.1 39.2 43.0 41.6 58.9 42.1 26.7
< 0.0001 0.0006
1 2.5 (0.8-8.1) 4.0 (1.2-13.1) NS
0.0012
NS
—
0.0312
NS
—
0.8327
NS
—
1 1.6 (1.1-2.3) 2.0 (1.4-2.9)
0.0192 < 0.0001
Synchronous (222) Metachronous (190) < 5 cm (332) Z5 cm (80) < 5 (343) Z5 (69)
40.6 43.5 43.5 35.0 43.1 35.3
0.0912
NS
—
0.0250
1 1.5 (1.1-2.0) NS
0.0213 —
Absence (381) Presence (31)
44.7 7.2
< 0.0001
1 2.0 (1.3-3.0)
0.0018
r5.0 mg/L (128) > 5.0 mg/L (251)
48.1 39.5
0.0573
< 0.0001
0.0518
P
0.1236 0.0232 —
Extrahepatic disease CEA*
NS
—
*Among 379 patients for whom the preoperative CEA level was known. CEA indicates carcinoembryonic antigen at hepatectomy; CI, confidence interval; HR, hazard ratio; NS, not selected.
1384 | www.ajsp.com
r
2014 Lippincott Williams & Wilkins
Am J Surg Pathol
Volume 38, Number 10, October 2014
cancer15,20; however, it is morphologically diverse according to individual tumors, possibly reflecting the wide spectrum of CAF functions. Previously, we proposed a simple 3-tiered histologic categorization of DR in the reactive fibrous zone on the basis of a histologic review and prognostic analyses of rectal cancer cases treated at St Marks Hospital.17,18 As yardsticks of criteria for the system, 2 distinctive histologic features of DR were adopted, that is, keloid-like collagen and myxoid stroma, both of which can be diagnosed on hematoxylin and eosin–stained slides. Principally, both stromal components do not appear in noninvasive tumors and are observed at the leading edge of invasive colorectal cancer wherein dynamic interaction between tumor cells and host cells, mediated by cytokines, chemokines, or growth factors, is embodied as a form of ECM remodeling. Keloid-like collagen is a term used for thick bands of hyalinized collagen in the reactive fibrous zone of colorectal cancer, which are morphologically similar to those in keloids–thick scars of the human skin or cornea, produced by the deposition of excessive amounts of collagen over prolonged periods.21 Pathophysiologically, a keloid is characterized by fibroblast proliferation and the synthesis of ECM components through the overexpression of growth factors, such as transforming growth factor b.22,23 The distinctive appearance of keloidlike collagens, which are immunohistochemically characterized by an excessive deposition of type 1 collagen,17 indicates the deviant activation of CAFs in the local tumor microenvironment. Myxoid stroma is formed as a result of extensive ECM remodeling and is characterized by a massive deposition of nonfibrillar ECM. Stromal cells, including myofibroblasts and bone marrow–derived cells, play a crucial role in ECM remodeling as sources of various types of protease activity, including matrix metalloproteases, cathepsins, and plasminogen activators.13,24 It is increasingly appreciated that ECM remodeling by such protease activity facilitates tumor invasion and metastasis by cleaving cell-adhesion molecules, by paving a path through the degradation of ECM, and by activating growth factors and cytokines.13,19 Myxoid stroma is also characterized by inconspicuous lymphocyte infiltration,17,18 indicating that the area is in a state of immunosuppression and may favor tumor progression. The St Marks Hospital series analyzed 862 curatively resected rectal cancer cases and found that the 5year survival rate was highest in the mature group (80%), followed by the intermediate (55%) and immature groups (27%).18 In the present study, we observed consistent results in both the independent institutions in terms of the adverse prognostic impact of the unfavorable DR pattern in the primary tumor: the rate of survival after hepatectomy was the highest in the mature group and lowest in the immature group. The significant prognostic impact was similarly observed in subset analyses for the primary tumor location, number of liver metastases, and, notably, liver metastasis timing. Multivariate analyses in the r
2014 Lippincott Williams & Wilkins
Primary Tumor Desmoplasia in CRLM
combined data set revealed that DR categorization was associated with prognostic outcome, independent of conventional prognostic parameters reflecting the anatomic extent of the disease in CRLM patients, such as the number and size of metastatic tumors and extrahepatic disease. It is becoming evident that characterizing the tumor microenvironment can provide prognostic information14; however, this is the first study to demonstrate that characterizing the microenvironment on the basis of the histologic component of ECMs, produced by CAFs as the result of the tumor-host interface, can predict oncological outcome in CRLM cases. Our study demonstrated that DR categorization was associated with the incidence of extrahepatic recurrence, particularly in the lung, peritoneum, and local region in the primary tumor, rather than recurrence in the remnant liver. The significance of pursuing prognostic markers in the primary tumors in cases of metastatic colorectal cancers is largely undetermined. However, the proportion of patients with recurrence limited to the remnant liver is only 30% to 40% in patients with recurrence after liver resection,25 and patient prognosis dominantly depends on the occurrence of extrahepatic metastasis originating from the primary tumor. It is therefore reasonable to assume that there are prognostic determinants in primary tumors, even in the CRLM setting. The molecular mechanisms and clinical advantages of histologic evaluation of the primary tumor microenvironment in CRLM cases remain to be elucidated in detail. Reportedly, DR categorization is significantly associated with the mismatch-repair status26; however, details on the genetic background of morphologic alterations of DR in colorectal cancers remain to be clarified. Therefore, this histologic categorization method can be used as a simple methodology to understand the status of the cancer microenvironment. Furthermore, recent studies have indicated that the microenvironment can affect the sensitivity of malignant cells to chemotherapy: tumor cells became less sensitive to chemotherapy following coculture with CAFs27,28 or CAF-derived ECM proteins.29 In addition, the accumulation of ECM components conceivably compresses blood and lymphatic vessels, leading to reduced perfusion and ultimately impeding the delivery of drugs to neoplastic cells.30 Further studies should focus on the correlation between histologic DR categorization and drug sensitivity and on whether metastatic tumors form DR patterns similar to those in the primary site. In conclusion, we found that the histologic characterization of the primary tumor microenvironment on the basis of products of CAFs helps in evaluating prognostic outcomes after hepatectomy in CRLM patients. The oncological benefit of neoadjuvant chemotherapy in patients suitable for curative hepatectomy remains debatable.31–33 The information obtained by comprehensively evaluating the microenvironmental regulation of metastases demonstrated in the primary tumor and the anatomical extent of the disease may allow us to determine the risk of www.ajsp.com |
1385
Ueno et al
Am J Surg Pathol
relapse after hepatectomy and aid in optimal patient selection for such a controversial treatment. REFERENCES 1. Sasaki A, Iwashita Y, Shibata K, et al. Analysis of preoperative prognostic factors for long-term survival after hepatic resection of liver metastasis of colorectal carcinoma. J Gastrointest Surg. 2005;9:374–380. 2. Minagawa M, Yamamoto J, Kosuge T, et al. Simplified staging system for predicting the prognosis of patients with resectable liver metastasis. Arch Surg. 2007;142:269–276. 3. Lee W-S, Kim MJ, Yun SH, et al. Risk factor stratification after simultaneous liver and colorectal resection for synchronous colorectal metastasis. Langenbecks Arch Surg. 2008;393:13–19. 4. Merkel S, Bialecki D, Meyer T, et al. Comparison of clinical risk scores predicting prognosis after resection of colorectal liver metastases. J Surg Oncol. 2009;100:349–357. 5. Beppu T, Sakamoto Y, Hasegawa K, et al. A nomogram predicting disease-free survival in patients with colorectal liver metastases treated with hepatic resection: multicenter data collection as a Project Study for Hepatic Surgery of the Japanese Society of Hepato-Biliary-Pancreatic Surgery. J Hepatobiliary Pancreat Sci. 2012;19:72–84. 6. Tanaka K, Shimada H, Ueda M, et al. Long-term characteristics of 5-year survivors after liver resection for colorectal metastases. Ann Surg Oncol. 2007;14:1336–1346. 7. Tan MCB, Castaldo ET, Gao F, et al. A prognostic system applicable to patients with resectable liver metastasis from colorectal carcinoma staged by positron emission tomography with [18F]fluoro2-deoxy-D-glucose: role of primary tumor variables. J Am Coll Surg. 2008;206:857–869. 8. Nordlinger B, Guiguet M, Vaillant J-C, et al. Surgical resection of colorectal carcinoma metastases to the liver: a prognostic scoring system to improve case selection, based on 1568 patients. Cancer. 1996;77:1254–1262. 9. Fong Y, Fortner J, Sun RL, et al. Clinical score for predicting recurrence after hepatic resection for metastatic colorectal cancer: analysis of 1001 consecutive cases. Ann Surg. 1999;230:309–321. 10. Gomez D, Cameron IC. Prognostic scores for colorectal liver metastasis: clinically important or an academic exercise? HPB (Oxford). 2010;12:227–238. 11. Ueno H, Mochizuki H, Hatsuse K, et al. Indicators for treatment strategies of colorectal liver metastases. Ann Surg. 2000;231:59–66. 12. National Comprehensive Cancer Network. NCCN clinical practice guidelines in oncology-colon cancer (version 3. 2014) 2012. Available at: http://www.nccn.org/professionals/physician_gls/pdf/ colon.pdf. Accessed February 1, 2014. 13. Joyce JA, Pollard JW. Microenvironmental regulation of metastasis. Nat Rev Cancer. 2009;9:239–252. 14. Allen M, Jones L. Jekyll and Hyde: the role of the microenvironment on the progression of cancer. J Pathol. 2011;223:162–176. 15. McAllister SS, Weinberg RA. Tumor-host interactions: a farreaching relationship. J Clin Oncol. 2010;28:4022–4028.
1386 | www.ajsp.com
Volume 38, Number 10, October 2014
16. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–674. 17. Ueno H, Jones A, Jass JR, et al. Clinicopathological significance of the ‘keloid-like’ collagen and myxoid stroma in advanced rectal cancer. Histopathology. 2002;40:327–334. 18. Ueno H, Jones AM, Wilkinson KH, et al. Histological categorisation of fibrotic cancer stroma in advanced rectal cancer. Gut. 2004;53:581–586. 19. Pietras K, O¨stman A. Hallmarks of cancer: interactions with the tumor stroma. Exp Cell Res. 2010;316:1324–1331. 20. Hewitt RE, Powe DG, Carter GI, et al. Desmoplasia and its relevance to colorectal tumour invasion. Int J Cancer. 1993;53: 62–69. 21. O’Sullivan ST, O’Connor TPF, O’Shaughnessy M. Aetiology and management of hypertrophic scars and keloids. Ann R Coll Surg Engl. 1996;78:168–175. 22. Bran GM, Goessler UR, Hormann K, et al. Keloids: current concepts of pathogenesis (review). Int J Mol Med. 2009;24:283–293. 23. Robles DT, Moore E, Draznin M, et al. Keloids: pathophysiology and management. Dermatol Online J. 2007;13:9. 24. Desmoulie`re A, Guyot C, Gabbiani G. The stroma reaction myofibroblast: a key player in the control of tumor cell behavior. Int J Dev Biol. 2004;48:509–517. 25. Ueno H, Mochizuki H, Hashiguchi Y, et al. Predictors of extrahepatic recurrence after resection of colorectal liver metastases. Br J Surg. 2004;53:581–586. 26. Wright C, Stewart I. Histopathology and mismatch repair status of 458 consecutive colorectal carcinomas. Am J Surg Pathol. 2003;27:1393–1406. 27. Mu¨erko¨ster S, Wegehenkel K, Arlt A, et al. Tumor stroma interactions induce chemoresistance in pancreatic ductal carcinoma cells involving increased secretion and paracrine effects of nitric oxide and interleukin-1beta. Cancer Res. 2004;64:1331–1337. 28. Shekhar MPV, Santner S, Carolin KA, et al. Direct involvement of breast tumor fibroblasts in the modulation of tamoxifen sensitivity. Am J Pathol. 2007;170:1546–1560. 29. Ahmed AA, Mills AD, Ibrahim AEK, et al. The extracellular matrix protein TGFBI induces microtubule stabilization and sensitizes ovarian cancers to paclitaxel. Cancer Cell. 2007;12:514–527. 30. Feig C, Gopinathan A, Neesse A, et al. The pancreas cancer microenvironment. Clin Cancer Res. 2012;18:4266–4276. 31. Nordlinger B, Sorbye H, Glimelius B, et al. Perioperative chemotherapy with FOLFOX4 and surgery versus surgery alone for resectable liver metastases from colorectal cancer (EORTC Intergroup trial 40983): a randomised controlled trial. Lancet. 2008;371: 1007–1016. 32. Adam R, Bhangui P, Poston G, et al. Is perioperative chemotherapy useful for solitary, metachronous, colorectal liver metastases? Ann Surg. 2010;252:774–787. 33. Chua TC, Saxena A, Liauw W, et al. Systematic review of randomized and nonrandomized trials of the clinical response and outcomes of neoadjuvant systematic chemotherapy for resectable colorectal liver metastases. Ann Surg Oncol. 2010;17:492–501.
r
2014 Lippincott Williams & Wilkins
