ORIGINAL CONTRIBUTION
Relative Value of Restaging MRI, CT, and FDG-PET Scan After Preoperative Chemoradiation for Rectal Cancer Daniel A. Schneider, M.D.1 • Timothy J. Akhurst, M.D., F.R.A.C.R.2 Samuel Y. Ngan, M.B.B.S., F.R.C.S.Edinb., F.R.A.C.R.3 Satish K. Warrier, M.B.B.S., M.S., F.R.A.C.S.1 • Michael Michael, M.D., F.R.A.C.P.4 Andrew C. Lynch, M.B.B.S., M.Med.Sci., F.R.A.C.S.1 • Luc Te Marvelde, Ph.D.5 Alexander G. Heriot, M.D., M.B.A., F.R.A.C.S., F.R.C.S.1 1 Department of Cancer Surgery, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 2 Centre for Molecular Imaging, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 3 Department of Radiation Oncology, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 4 Department of Cancer Medicine, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 5 Centre for Biostatistics and Clinical Trials, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia
BACKGROUND: Management of rectal cancer has become multidisciplinary and is driven by the stage of the disease, with increased focus on restaging rectal cancer after neoadjuvant therapy.
MAIN OUTCOME MEASURES: Comparisons of the tumor stage among different imaging modalities before and after neoadjuvant chemoradiation were performed. The impact of restaging on the management plan was assessed.
OBJECTIVE: The purpose of this study was to assess the relative impact of restaging after preoperative chemoradiation with FDG-PET scan, CT, and MRI in the management of patients with rectal cancer.
RESULTS: The stage at presentation was T2, 8.04%; T3, 65.33%; T4, 26.63%; N0, 17.09%; N1, 47.74%; N2, 34.67%; M0, 81.91%; and M1, 18.09%. Changes in disease stage postneoadjuvant chemoradiation were observed in 99 patients (50%). The management plans of 29 patients (15%) were changed. The impact of each restaging modality on management for all of the patients was positron emission tomography, 11%; CT, 4%; and MRI, 4%. In patients with metastatic disease at primary staging, the relative impact of each restaging modality in changing management was positron emission tomography, 32%; CT, 18%; and MRI, 6%.
DESIGN: This was a retrospective study from a single institution. SETTINGS: This study was conducted at a tertiary cancer
center. PATIENTS: A total of 199 patients met the inclusion criteria: patients with rectal adenocarcinoma; staged with positron emission tomography, CT, and MRI; T2 to T4, N0 to N2, M0 to M1; treated with neoadjuvant chemoradiation 50.4 Gy and infusional 5-fluorouracil; and restaged 4 weeks after chemoradiation before surgery between 2003 and 2013.
Financial Disclosure: None reported. Podium presentation at the meeting of The American Society of Colon and Rectal Surgeons, Boston, MA, May 30 to June 3, 2015.
LIMITATIONS: This study was limited by its single-center and retrospective design. Operations were performed 4 weeks after restaging. CONCLUSIONS: Changes in the extent of disease after longcourse chemoradiotherapy result in changes of management in a significant percentage of patients. Positron emission tomography has the most significant impact in the change of management overall, and its use in restaging advanced rectal cancer should be further explored.
Correspondence: Alexander G. Heriot, M.D., M.B.A., F.R.A.C.S., F.R.C.S., Department of Cancer Surgery, Peter MacCallum Cancer Centre, Locked Bag 1 A’Beckett St, East Melbourne, Victoria 8006, Australia. E-mail: [email protected]
KEY WORDS: Chemoradiotherapy; Neoadjuvant therapy; Rectal cancer; Restaging.
Dis Colon Rectum 2016; 59: 179–186 DOI: 10.1097/DCR.0000000000000557 © The ASCRS 2015
T
Diseases of the Colon & Rectum Volume 59: 3 (2016)
he management of rectal cancer has become multidisciplinary and is driven by the stage of the disease. Preoperative neoadjuvant therapy, usually 179
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
180
chemoradiotherapy, reduces the risk of local recurrence in locally advanced rectal cancer,1 and the application is determined through preoperative imaging with pelvic MRI2,3 and/or endoanal ultrasound.4 Distant staging is important because it directly influences management. Limited metastatic disease can be resected for cure through liver, pulmonary, or peritoneal resection. Extensive metastatic disease may result in a palliative approach, often involving chemotherapy alone. Both CT and positron emission tomography (PET)-CT have demonstrated impact on staging and management of rectal cancer. There has been increased focus on restaging rectal cancer after the application of neoadjuvant therapy.5,6 Downstaging of local disease may result in a change in the extent of surgical resection7,8 or potentially even avoidance of surgery and a watch-and-wait approach if it is considered that the tumor has undergone a complete pathological response.9 Metabolic response of the primary tumor on PET-CT may be prognostic of long-term outcome,10 as may the presence of extramural venous invasion on restaging MRI.11 Changes in metastatic disease may alter the management approach and intent. The aim of the study was to assess the change in disease stage observed with PET-CT, CT, and MRI before and after neoadjuvant chemoradiotherapy in patients with rectal cancer. The study also sought to examine the relative impact of each staging modality by assessing which imaging modality altered patient management.
PATIENTS AND METHODS This retrospective study was approved by the internal ethics review board and was performed in accordance with the Helsinki Declaration of 1975 as revised in 2000. Eligibility criteria included patients with histologically confirmed adenocarcinoma of the rectum, within 15 cm of the anal verge, and disease stage T2 to T4/N0 to N2/M0 to M1 who received neoadjuvant long-course chemoradiation (CRT; LCCRT) staged with PET, CT, and MRI before and after LCCRT and treated at our institute between 2003 and 2013. The standard protocol for patients presenting with newly diagnosed rectal cancer was complete staging with sigmoidoscopy, rectal biopsy, colonoscopy, pelvic rectal MRI for T and N stage, CT chest-abdomen-pelvis and FDG-PET for assessment of local and possible distant disease. Each patient had their imaging and treatment planning performed at our center. Each patient’s staging imaging was reviewed at the multidisciplinary team meeting (MDT) to formulate a treatment plan. The treatment intent (either curative or palliative) and management plan were documented in the patient’s medical chart. Treatment intent was determined by the extent of patient disease in combination with the likely prognosis and was a consensus decision arrived at by the combined
Schneider et al: Rectal Cancer Restaging After Chemoradiation
MDT. Management plan for the patient was the exact plan for treatment, combining both details of CRT and surgical approaches and other interventions. Members of this multidisciplinary team consisted of colorectal surgeons, radiation oncologists, medical oncologists, radiologists, nuclear medicine physicians, specialist nurses, and allied health workers. Patients with MRI staged T3 to T4, highrisk T2 tumors with close proximity to the anal verge, or N+ were treated with LCCRT. Four weeks after completion of LCCRT, patients were restaged with MRI, CT, and PET. The restaging imaging and management plan of each patient was reviewed. The treatment intent and management plan from the second MDT was documented in the patient’s medical chart. Patients with metastatic disease were considered curative or palliative on treatment intent. All of the patients in the study received LCCRT, with our standard protocol stated below. Some patients were given a sandwich protocol,12 which involved additional chemotherapy. If metastatic disease was considered resectable, it was resected. The timing and nature of metastectomy were decided based on the response to therapy and individualized for each patient. The standard LCCRT consisted of pelvic radiotherapy 50.4 Gy in 1.8-Gy per fraction increments over 5.5 weeks, with concurrent infusion of 5-fluorouracil at 225 mg/m2 per day. Surgery was to be performed 4 to 6 weeks after the completion of LCCRT. Patients were to receive adjuvant chemotherapy after recovery from surgery. Eleven patients (5.5%) treated according to a clinical trial protocol were included.12 The protocol of the PET was described in our previous study from our center.10 Patients were studied on a dedicated PET/CT scanner (Discovery 690, Discovery DTSE, GE Healthcare, Little Chalfont, United Kingdom; or Siemens Biograph PET CT system, Siemens Healthcare USA, Malvern, PA) 1 hour after injection of 300 to 400 MBq of 18F-FDG. Images were acquired from the neck to the upper thigh. A retrospective chart review was undertaken to identify patients who met the inclusion criteria. The results of staging and restaging investigations were recorded, including dates of investigations and TNM stages. Staging and restaging MDT notes were reviewed and the treatment intent and management plan from each MDT was recorded, including the regimen of CRT that each patient received. Time to restaging was calculated from the date of completion of CRT to the date of restaging investigations. Progression of disease between staging and restaging was classified as progression of local disease, progression of existing metastatic disease, or development of new metastatic disease. A change in stage was defined as a change in TNM stage grouping. In the case of metastatic disease at primary staging, this was downstaged if there was complete response on imaging and upstaged if there was
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
181
Diseases of the Colon & Rectum Volume 59: 3 (2016)
TABLE 1. Change in stage after neoadjuvant LCCRT Variable Downstage No change Upstage Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
81 (51% (43%–59%)) 71 (44% (37%–52%)) 8 (5% (2%–10%)) 3
9 (25% (12%–42%)) 26 (72% (55%–86%)) 1 (3% (0%–15%)) 0
90 (46% (39%–53%)) 97 (49% (42%–57%)) 9 (5% (2%–9%)) 3
LCCRT = long-course chemoradiation.
evelopment of new metastatic disease. Changes in stage d and changes in management were attributed to each imaging modality that was primarily responsible for a change in disease stage or management. For the purposes of validation, histological reports were reviewed and TNM staging was recorded. Dates of surgery were recorded. Patient demographics, staging and restaging, and management details were summarized using descriptive statistics including counts; frequencies for categorical variables; and mean, SD, median, and range for continuous variables. The proportion of patients for whom staging/management differed between restaging scans and pathology was expressed as the proportion with its associated exact 95% CI based on the binomial distribution. All of the analyses were performed in R (version 3.0.1).
RESULTS A total of 199 patients (137 men and 62 women) were included in the study. The median age at diagnosis was 60.7 years (range, 30.2–88.0 years). The median time period between the completion of CRT and restaging was 29.3 days (range, 9–56 days). Tumor stages after primary staging were stage I, 3% (n = 5); stage IIA, 10% (n = 19); stage IIC, 3% (n = 5); stage IIIA, 5% (n = 9); stage IIIB, 52% (n = 103); stage IIIC, 11% (n = 22); and stage IV, 18% (n = 36). The 5 patients with stage I disease were included on account of
having very low T2 tumors, near the anal verge, and were given LCCRT with the aim of being able to perform a less extensive resection. Treatment intent was curative in 97.0% (n = 193), palliative in 2.5% (n = 5), and undocumented in 1 patient. Change in Stage at Restaging
After LCCRT and restaging, 99 patients (50%) had a change in stage with disease downstaged in 90 patients. Stage IV disease was noted in 35 patients (18%) at restaging compared with 36 patients at primary staging (Tables 1–4). Restaging imaging demonstrated disease progression in 21 patients (11%). Of the patients who had no metastatic disease at staging, 12 (7%) had disease progression (local = 4, metastasis = 8, including 1 both local and distant). Of the 36 patients who presented with distant metastatic disease, 9 (25%) had progression, (local = 3, metastasis progression = 8, new metastasis = 4). Three patients had a negative CT and positive PET scan for metastatic disease. In 56% of patients (95% CI, 48%–63%), combined imaging overestimated staging compared with staging based on pathology, whereas in 18% of patients (95% CI, 13%–24%), combined imaging underestimated staging compared with staging based on pathology. When broken down to individual imaging modalities and their comparison with histological data locally, locoregionally, and distantly, accuracy of restaging imaging was improved compared with the cumulative restaging accuracy (Tables 3 and 4). T stage was accurate on restaging MRI in
TABLE 2. Change in stage after neoadjuvant LCCRT: responsible imaging modalities Variable PET No Yes Missing MRI No Yes Missing CT No Yes Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
45 (51% (40%–61%)) 44 (49% (39%–60%)) 0
1 (10% (0%–45%)) 9 (90% (55%–100%)) 0
46 (46% (36%–57%)) 53 (54% (43%–64%)) 3
12 (13% (7%–22%)) 77 (87% (78%–93%)) 0
6 (60% (26%–88%)) 4 (40% (12%–74%)) 0
18 (18% (11%–27%)) 81 (82% (73%–89%)) 3
73 (82% (72%–89%)) 16 (18% (11%–28%)) 0
7 (70% (35%–93%)) 3 (30% (7%–65%)) 0
80 (81% (72%–88%)) 19 (19% (12%–28%)) 3
LCCRT = long-course chemoradiation; PET = positron emission tomography.
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
182
Schneider et al: Rectal Cancer Restaging After Chemoradiation
TABLE 3. Accuracy of restaging imaging compared with histology Variable
Accurate, n (%)
MRI, T stage MRI, N stage PET, N stage CT, N stage
Overstaged, n (%)
84 (43) 112 (56) 119 (60) 111 (56)
Understaged, n (%)
69 (34) 48 (24) 16 (8) 12 (6)
Missing data, n (%)
35 (18) 29 (15) 54 (27) 31 (15)
11 (5) 10 (5) 10 (5) 45 (23)
PET = positron emission tomography.
43% of patients, with 34% of patients overstaged and 18% understaged. Meanwhile, the range of N stage accuracy across MRI, PET scan, and CT was 56% to 60%, with MRI shown to overstage more than understage and PET scan to understage more than overstage. In patients who had resections of metastatic disease, the accuracy ranged between 54% and 80%. This is a small patient group, because not all of the patients with stage IV disease had metastases resected. Four patients with metastasectomies had their management changed as a result of restaging.
Impact of Restaging Imaging by Modality
Change in Treatment Intent at Restaging
The treatment intent was changed after LCCRT in 4 patients (2%; Fig. 1). Three patients had metastatic disease at primary staging. Treatment intent was changed from curative to palliative in 3 and from palliative to curative in 1 patient. Despite 21 patients having progression of disease, as mentioned above, the majority were treated with curative intent whereby resections of metastases were planned, for example, concurrent resection of hepatic metastases. As such, these patients were still treated with curative intent. Change in Management Plan at Restaging
A change in management plan occurred in 29 patients after restaging (15%), with a total of 48 events of change in management amongst these 29 patients (Tables 5–9). Twelve of the 36 patients who presented with metastatic disease at staging had a change in management, composing 33% of this subset of patients. Meanwhile, of the 163 patients who presented at staging with stage I to III disease, 17 had management plan changed, composed of 10% of this subset of patients. It is worth noting that, whereas 50% of patients had a change in stage at restaging, only 15% had a change in management. As such, some changes in stage are irrelevant to therapeutic planning and should be interpreted carefully. The most common change in management plan was a change in surgical plan (Table 8), which occurred in TABLE 4. Accuracy of restaging imaging compared with histology for patients with metastatic disease with resection of metastatic disease Variable PET, M stage CT, M stage
Accurate, n
Overstaged, n
7 8
2 0
PET = positron emission tomography.
18 patients, including 4 patients with distant metastatic disease at staging. Changes in surgical plan included 7 patients who required more extensive resections of either local or distant disease and 5 patients who required new intraoperative radiotherapy (Table 9). Eight patients had their management plans changed to manage metastatic disease (new metastasis = 5, metastasis progression = 3). After restaging, 7 patients overall (4%) were given further chemotherapy that had not been part of their management plan at staging.
Understaged, n 4 2
Change in Stage
PET was responsible for a change in stage in 53 patients overall (27%), which was reflected in equal proportions among both patients who did not have distant metastatic disease at staging (27%) and those who did (26%); (Tables 1 and 2). MRI was responsible for a change in stage in 81 patients overall (41%), which was a more prominent finding in patients who did not have distant metastatic disease at staging (48%) compared with those with distant metastatic disease at staging, where it was responsible for a change in stage in only 4 patients (12%). CT, either performed in a combined PETCT scan or as a stand-alone investigation, was responsible for a change in stage in 10% of patients overall, which were distributed in equal proportions between patients without distant metastatic disease at staging (10%) and those who did have distant metastatic disease at staging (9%). Change in Management
PET was responsible for a change in management in 22 patients overall (11%; Tables 6 and 7). The resulting change in management was more prevalent in patients who had distant metastatic disease at staging than without (32% versus 7%). MRI was responsible for a change in management in 8 patients overall (4%), represented in equal proportions between those with metastatic disease and no metastasis at staging. CT was responsible for a change in management in 8 patients overall (4%), with a greater impact in patients with metastatic disease at staging (18% versus 1%).
DISCUSSION Restaging after CRT in advanced rectal cancer altered treatment plans in 15% of patients, with an alteration in surgical strategy comprising the most common change (62%
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
183
Diseases of the Colon & Rectum Volume 59: 3 (2016)
All patients (n = 199) Staging
Treatment intent = curative (n = 193) M0 (n = 162) M1 (n = 31)
Treatment intent = palliative (n = 5) M0 (n = 1) M1 (n = 4)
LCCRT
LCCRT
Treatment intent = curative (n = 190) M0 (n = 161) M1 (n = 29)
Treatment intent = palliative (n = 3) M0 (n = 1) M1 (n = 2)
Treatment intent = curative (n = 1) M0 (n = 0) M1 (n = 1)
Not documented (n = 1)
Treatment intent = palliative (n = 4) M0 (n = 1) M1 (n = 3)
FIGURE 1. Consort diagram regarding treatment intent for all patients before and after long-course chemoradiation (LCCRT).
of patients with a change in management). The use of PET scan and MRI restaging are highlighted by the current study, with PET-CT scan altering distant restaging more, whereas MRI restaging impacts local restaging. In addition, notable positive findings from the study included 7 patients where additional chemotherapy was administered before planned surgery, 8 patients who underwent more extensive surgical resections, and 5 additional patients in whom intraoperative radiation was offered after restaging. The study strengths include a standardized institutional protocol whereby all of the patients were restaged with all modalities at the same time interval (4 weeks), with prospective documentation of tumor board meetings to guide management. Our study is primarily limited by its single-center, retrospective design. In addition, given that the records were reviewed retrospectively, a few cases did lack clear management and operative details. The imaging was also reported by numerous radiologists and nuclear medicine technicians over the time period. However, we feel that the
conclusions represent a real-world experience of restaging as a routine component of the clinical management of rectal cancer. Another potential critique is the timing of the restaging. In line with the current literature, as an institution we have extended the interval from cessation of CRT to definitive surgery. A logical question would be whether we are restaging too early. Restaging was undertaken at 4 weeks post-LCCRT. Surgery was initially done at 8 weeks and eventually lengthened to 10 weeks in accordance with the literature. We agree that this leads to a lower accuracy rate compared with histological staging given the time interval. The department has addressed this, and timing of restaging is being reviewed. Restaging local disease remains a challenge, because there is variable accuracy. In our series, there was an accuracy of 43% for T stage and 56% to 60% for N stage as seen in Table 3. However, local staging does not change management in most cases, and with the exception of large tumors or those that threaten resection margins, management of local disease is often guided by initial staging. Conversely,
TABLE 5. Change in management plan after neoadjuvant LCCRT Variable
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
No Yes Missing
145 (90% (84%–94%)) 17 (10% (6%–16%)) 1
22 (65% (46%–80%)) 12 (35% (20%–54%)) 2
All patients (N = 199), n (OR (95% CI)) 167 (85% (79%–90%)) 29 (15% (10%–21%)) 3
LCCRT = long-course chemoradiation.
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
184
Schneider et al: Rectal Cancer Restaging After Chemoradiation
TABLE 6. Change in management plan after neoadjuvant LCCRT: responsible imaging modalities Variable PET No Yes Missing MRI No Yes Missing CT No Yes Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
151 (93% (88%–97%)) 11 (7% (3%–12%)) 1
23 (68% (49%–83%)) 11 (32% (17%–51%)) 2
174 (89% (84%–93%)) 22 (11% (7%–16%)) 3
156 (96% (92%–99%)) 6 (4% (1%–8%)) 1
32 (94% (80%–99%)) 2 (6% (1%–20%)) 2
188 (96% (92%–98%)) 8 (4% (2%–8%)) 3
160 (99% (96%–100%)) 2 (1% (0%–4%)) 1
28 (82% (65%–93%)) 6 (18% (7%–35%)) 2
188 (96% (92%–98%)) 8 (4% (2%–8%)) 3
LCCRT = long-course chemoradiation; PET = positron emission tomography.
staging of distant disease is more accurate. Although we had a limited number of patients who had a metastasectomy to assess the accuracy of restaging, given that not all metastatic disease identified was resectable, the M-stage accuracy was
Relative Value of Restaging MRI, CT, and FDG-PET Scan After Preoperative Chemoradiation for Rectal Cancer Daniel A. Schneider, M.D.1 • Timothy J. Akhurst, M.D., F.R.A.C.R.2 Samuel Y. Ngan, M.B.B.S., F.R.C.S.Edinb., F.R.A.C.R.3 Satish K. Warrier, M.B.B.S., M.S., F.R.A.C.S.1 • Michael Michael, M.D., F.R.A.C.P.4 Andrew C. Lynch, M.B.B.S., M.Med.Sci., F.R.A.C.S.1 • Luc Te Marvelde, Ph.D.5 Alexander G. Heriot, M.D., M.B.A., F.R.A.C.S., F.R.C.S.1 1 Department of Cancer Surgery, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 2 Centre for Molecular Imaging, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 3 Department of Radiation Oncology, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 4 Department of Cancer Medicine, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia 5 Centre for Biostatistics and Clinical Trials, Peter MacCallum Cancer Centre, East Melbourne, Victoria, Australia
BACKGROUND: Management of rectal cancer has become multidisciplinary and is driven by the stage of the disease, with increased focus on restaging rectal cancer after neoadjuvant therapy.
MAIN OUTCOME MEASURES: Comparisons of the tumor stage among different imaging modalities before and after neoadjuvant chemoradiation were performed. The impact of restaging on the management plan was assessed.
OBJECTIVE: The purpose of this study was to assess the relative impact of restaging after preoperative chemoradiation with FDG-PET scan, CT, and MRI in the management of patients with rectal cancer.
RESULTS: The stage at presentation was T2, 8.04%; T3, 65.33%; T4, 26.63%; N0, 17.09%; N1, 47.74%; N2, 34.67%; M0, 81.91%; and M1, 18.09%. Changes in disease stage postneoadjuvant chemoradiation were observed in 99 patients (50%). The management plans of 29 patients (15%) were changed. The impact of each restaging modality on management for all of the patients was positron emission tomography, 11%; CT, 4%; and MRI, 4%. In patients with metastatic disease at primary staging, the relative impact of each restaging modality in changing management was positron emission tomography, 32%; CT, 18%; and MRI, 6%.
DESIGN: This was a retrospective study from a single institution. SETTINGS: This study was conducted at a tertiary cancer
center. PATIENTS: A total of 199 patients met the inclusion criteria: patients with rectal adenocarcinoma; staged with positron emission tomography, CT, and MRI; T2 to T4, N0 to N2, M0 to M1; treated with neoadjuvant chemoradiation 50.4 Gy and infusional 5-fluorouracil; and restaged 4 weeks after chemoradiation before surgery between 2003 and 2013.
Financial Disclosure: None reported. Podium presentation at the meeting of The American Society of Colon and Rectal Surgeons, Boston, MA, May 30 to June 3, 2015.
LIMITATIONS: This study was limited by its single-center and retrospective design. Operations were performed 4 weeks after restaging. CONCLUSIONS: Changes in the extent of disease after longcourse chemoradiotherapy result in changes of management in a significant percentage of patients. Positron emission tomography has the most significant impact in the change of management overall, and its use in restaging advanced rectal cancer should be further explored.
Correspondence: Alexander G. Heriot, M.D., M.B.A., F.R.A.C.S., F.R.C.S., Department of Cancer Surgery, Peter MacCallum Cancer Centre, Locked Bag 1 A’Beckett St, East Melbourne, Victoria 8006, Australia. E-mail: [email protected]
KEY WORDS: Chemoradiotherapy; Neoadjuvant therapy; Rectal cancer; Restaging.
Dis Colon Rectum 2016; 59: 179–186 DOI: 10.1097/DCR.0000000000000557 © The ASCRS 2015
T
Diseases of the Colon & Rectum Volume 59: 3 (2016)
he management of rectal cancer has become multidisciplinary and is driven by the stage of the disease. Preoperative neoadjuvant therapy, usually 179
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
180
chemoradiotherapy, reduces the risk of local recurrence in locally advanced rectal cancer,1 and the application is determined through preoperative imaging with pelvic MRI2,3 and/or endoanal ultrasound.4 Distant staging is important because it directly influences management. Limited metastatic disease can be resected for cure through liver, pulmonary, or peritoneal resection. Extensive metastatic disease may result in a palliative approach, often involving chemotherapy alone. Both CT and positron emission tomography (PET)-CT have demonstrated impact on staging and management of rectal cancer. There has been increased focus on restaging rectal cancer after the application of neoadjuvant therapy.5,6 Downstaging of local disease may result in a change in the extent of surgical resection7,8 or potentially even avoidance of surgery and a watch-and-wait approach if it is considered that the tumor has undergone a complete pathological response.9 Metabolic response of the primary tumor on PET-CT may be prognostic of long-term outcome,10 as may the presence of extramural venous invasion on restaging MRI.11 Changes in metastatic disease may alter the management approach and intent. The aim of the study was to assess the change in disease stage observed with PET-CT, CT, and MRI before and after neoadjuvant chemoradiotherapy in patients with rectal cancer. The study also sought to examine the relative impact of each staging modality by assessing which imaging modality altered patient management.
PATIENTS AND METHODS This retrospective study was approved by the internal ethics review board and was performed in accordance with the Helsinki Declaration of 1975 as revised in 2000. Eligibility criteria included patients with histologically confirmed adenocarcinoma of the rectum, within 15 cm of the anal verge, and disease stage T2 to T4/N0 to N2/M0 to M1 who received neoadjuvant long-course chemoradiation (CRT; LCCRT) staged with PET, CT, and MRI before and after LCCRT and treated at our institute between 2003 and 2013. The standard protocol for patients presenting with newly diagnosed rectal cancer was complete staging with sigmoidoscopy, rectal biopsy, colonoscopy, pelvic rectal MRI for T and N stage, CT chest-abdomen-pelvis and FDG-PET for assessment of local and possible distant disease. Each patient had their imaging and treatment planning performed at our center. Each patient’s staging imaging was reviewed at the multidisciplinary team meeting (MDT) to formulate a treatment plan. The treatment intent (either curative or palliative) and management plan were documented in the patient’s medical chart. Treatment intent was determined by the extent of patient disease in combination with the likely prognosis and was a consensus decision arrived at by the combined
Schneider et al: Rectal Cancer Restaging After Chemoradiation
MDT. Management plan for the patient was the exact plan for treatment, combining both details of CRT and surgical approaches and other interventions. Members of this multidisciplinary team consisted of colorectal surgeons, radiation oncologists, medical oncologists, radiologists, nuclear medicine physicians, specialist nurses, and allied health workers. Patients with MRI staged T3 to T4, highrisk T2 tumors with close proximity to the anal verge, or N+ were treated with LCCRT. Four weeks after completion of LCCRT, patients were restaged with MRI, CT, and PET. The restaging imaging and management plan of each patient was reviewed. The treatment intent and management plan from the second MDT was documented in the patient’s medical chart. Patients with metastatic disease were considered curative or palliative on treatment intent. All of the patients in the study received LCCRT, with our standard protocol stated below. Some patients were given a sandwich protocol,12 which involved additional chemotherapy. If metastatic disease was considered resectable, it was resected. The timing and nature of metastectomy were decided based on the response to therapy and individualized for each patient. The standard LCCRT consisted of pelvic radiotherapy 50.4 Gy in 1.8-Gy per fraction increments over 5.5 weeks, with concurrent infusion of 5-fluorouracil at 225 mg/m2 per day. Surgery was to be performed 4 to 6 weeks after the completion of LCCRT. Patients were to receive adjuvant chemotherapy after recovery from surgery. Eleven patients (5.5%) treated according to a clinical trial protocol were included.12 The protocol of the PET was described in our previous study from our center.10 Patients were studied on a dedicated PET/CT scanner (Discovery 690, Discovery DTSE, GE Healthcare, Little Chalfont, United Kingdom; or Siemens Biograph PET CT system, Siemens Healthcare USA, Malvern, PA) 1 hour after injection of 300 to 400 MBq of 18F-FDG. Images were acquired from the neck to the upper thigh. A retrospective chart review was undertaken to identify patients who met the inclusion criteria. The results of staging and restaging investigations were recorded, including dates of investigations and TNM stages. Staging and restaging MDT notes were reviewed and the treatment intent and management plan from each MDT was recorded, including the regimen of CRT that each patient received. Time to restaging was calculated from the date of completion of CRT to the date of restaging investigations. Progression of disease between staging and restaging was classified as progression of local disease, progression of existing metastatic disease, or development of new metastatic disease. A change in stage was defined as a change in TNM stage grouping. In the case of metastatic disease at primary staging, this was downstaged if there was complete response on imaging and upstaged if there was
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
181
Diseases of the Colon & Rectum Volume 59: 3 (2016)
TABLE 1. Change in stage after neoadjuvant LCCRT Variable Downstage No change Upstage Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
81 (51% (43%–59%)) 71 (44% (37%–52%)) 8 (5% (2%–10%)) 3
9 (25% (12%–42%)) 26 (72% (55%–86%)) 1 (3% (0%–15%)) 0
90 (46% (39%–53%)) 97 (49% (42%–57%)) 9 (5% (2%–9%)) 3
LCCRT = long-course chemoradiation.
evelopment of new metastatic disease. Changes in stage d and changes in management were attributed to each imaging modality that was primarily responsible for a change in disease stage or management. For the purposes of validation, histological reports were reviewed and TNM staging was recorded. Dates of surgery were recorded. Patient demographics, staging and restaging, and management details were summarized using descriptive statistics including counts; frequencies for categorical variables; and mean, SD, median, and range for continuous variables. The proportion of patients for whom staging/management differed between restaging scans and pathology was expressed as the proportion with its associated exact 95% CI based on the binomial distribution. All of the analyses were performed in R (version 3.0.1).
RESULTS A total of 199 patients (137 men and 62 women) were included in the study. The median age at diagnosis was 60.7 years (range, 30.2–88.0 years). The median time period between the completion of CRT and restaging was 29.3 days (range, 9–56 days). Tumor stages after primary staging were stage I, 3% (n = 5); stage IIA, 10% (n = 19); stage IIC, 3% (n = 5); stage IIIA, 5% (n = 9); stage IIIB, 52% (n = 103); stage IIIC, 11% (n = 22); and stage IV, 18% (n = 36). The 5 patients with stage I disease were included on account of
having very low T2 tumors, near the anal verge, and were given LCCRT with the aim of being able to perform a less extensive resection. Treatment intent was curative in 97.0% (n = 193), palliative in 2.5% (n = 5), and undocumented in 1 patient. Change in Stage at Restaging
After LCCRT and restaging, 99 patients (50%) had a change in stage with disease downstaged in 90 patients. Stage IV disease was noted in 35 patients (18%) at restaging compared with 36 patients at primary staging (Tables 1–4). Restaging imaging demonstrated disease progression in 21 patients (11%). Of the patients who had no metastatic disease at staging, 12 (7%) had disease progression (local = 4, metastasis = 8, including 1 both local and distant). Of the 36 patients who presented with distant metastatic disease, 9 (25%) had progression, (local = 3, metastasis progression = 8, new metastasis = 4). Three patients had a negative CT and positive PET scan for metastatic disease. In 56% of patients (95% CI, 48%–63%), combined imaging overestimated staging compared with staging based on pathology, whereas in 18% of patients (95% CI, 13%–24%), combined imaging underestimated staging compared with staging based on pathology. When broken down to individual imaging modalities and their comparison with histological data locally, locoregionally, and distantly, accuracy of restaging imaging was improved compared with the cumulative restaging accuracy (Tables 3 and 4). T stage was accurate on restaging MRI in
TABLE 2. Change in stage after neoadjuvant LCCRT: responsible imaging modalities Variable PET No Yes Missing MRI No Yes Missing CT No Yes Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
45 (51% (40%–61%)) 44 (49% (39%–60%)) 0
1 (10% (0%–45%)) 9 (90% (55%–100%)) 0
46 (46% (36%–57%)) 53 (54% (43%–64%)) 3
12 (13% (7%–22%)) 77 (87% (78%–93%)) 0
6 (60% (26%–88%)) 4 (40% (12%–74%)) 0
18 (18% (11%–27%)) 81 (82% (73%–89%)) 3
73 (82% (72%–89%)) 16 (18% (11%–28%)) 0
7 (70% (35%–93%)) 3 (30% (7%–65%)) 0
80 (81% (72%–88%)) 19 (19% (12%–28%)) 3
LCCRT = long-course chemoradiation; PET = positron emission tomography.
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
182
Schneider et al: Rectal Cancer Restaging After Chemoradiation
TABLE 3. Accuracy of restaging imaging compared with histology Variable
Accurate, n (%)
MRI, T stage MRI, N stage PET, N stage CT, N stage
Overstaged, n (%)
84 (43) 112 (56) 119 (60) 111 (56)
Understaged, n (%)
69 (34) 48 (24) 16 (8) 12 (6)
Missing data, n (%)
35 (18) 29 (15) 54 (27) 31 (15)
11 (5) 10 (5) 10 (5) 45 (23)
PET = positron emission tomography.
43% of patients, with 34% of patients overstaged and 18% understaged. Meanwhile, the range of N stage accuracy across MRI, PET scan, and CT was 56% to 60%, with MRI shown to overstage more than understage and PET scan to understage more than overstage. In patients who had resections of metastatic disease, the accuracy ranged between 54% and 80%. This is a small patient group, because not all of the patients with stage IV disease had metastases resected. Four patients with metastasectomies had their management changed as a result of restaging.
Impact of Restaging Imaging by Modality
Change in Treatment Intent at Restaging
The treatment intent was changed after LCCRT in 4 patients (2%; Fig. 1). Three patients had metastatic disease at primary staging. Treatment intent was changed from curative to palliative in 3 and from palliative to curative in 1 patient. Despite 21 patients having progression of disease, as mentioned above, the majority were treated with curative intent whereby resections of metastases were planned, for example, concurrent resection of hepatic metastases. As such, these patients were still treated with curative intent. Change in Management Plan at Restaging
A change in management plan occurred in 29 patients after restaging (15%), with a total of 48 events of change in management amongst these 29 patients (Tables 5–9). Twelve of the 36 patients who presented with metastatic disease at staging had a change in management, composing 33% of this subset of patients. Meanwhile, of the 163 patients who presented at staging with stage I to III disease, 17 had management plan changed, composed of 10% of this subset of patients. It is worth noting that, whereas 50% of patients had a change in stage at restaging, only 15% had a change in management. As such, some changes in stage are irrelevant to therapeutic planning and should be interpreted carefully. The most common change in management plan was a change in surgical plan (Table 8), which occurred in TABLE 4. Accuracy of restaging imaging compared with histology for patients with metastatic disease with resection of metastatic disease Variable PET, M stage CT, M stage
Accurate, n
Overstaged, n
7 8
2 0
PET = positron emission tomography.
18 patients, including 4 patients with distant metastatic disease at staging. Changes in surgical plan included 7 patients who required more extensive resections of either local or distant disease and 5 patients who required new intraoperative radiotherapy (Table 9). Eight patients had their management plans changed to manage metastatic disease (new metastasis = 5, metastasis progression = 3). After restaging, 7 patients overall (4%) were given further chemotherapy that had not been part of their management plan at staging.
Understaged, n 4 2
Change in Stage
PET was responsible for a change in stage in 53 patients overall (27%), which was reflected in equal proportions among both patients who did not have distant metastatic disease at staging (27%) and those who did (26%); (Tables 1 and 2). MRI was responsible for a change in stage in 81 patients overall (41%), which was a more prominent finding in patients who did not have distant metastatic disease at staging (48%) compared with those with distant metastatic disease at staging, where it was responsible for a change in stage in only 4 patients (12%). CT, either performed in a combined PETCT scan or as a stand-alone investigation, was responsible for a change in stage in 10% of patients overall, which were distributed in equal proportions between patients without distant metastatic disease at staging (10%) and those who did have distant metastatic disease at staging (9%). Change in Management
PET was responsible for a change in management in 22 patients overall (11%; Tables 6 and 7). The resulting change in management was more prevalent in patients who had distant metastatic disease at staging than without (32% versus 7%). MRI was responsible for a change in management in 8 patients overall (4%), represented in equal proportions between those with metastatic disease and no metastasis at staging. CT was responsible for a change in management in 8 patients overall (4%), with a greater impact in patients with metastatic disease at staging (18% versus 1%).
DISCUSSION Restaging after CRT in advanced rectal cancer altered treatment plans in 15% of patients, with an alteration in surgical strategy comprising the most common change (62%
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
183
Diseases of the Colon & Rectum Volume 59: 3 (2016)
All patients (n = 199) Staging
Treatment intent = curative (n = 193) M0 (n = 162) M1 (n = 31)
Treatment intent = palliative (n = 5) M0 (n = 1) M1 (n = 4)
LCCRT
LCCRT
Treatment intent = curative (n = 190) M0 (n = 161) M1 (n = 29)
Treatment intent = palliative (n = 3) M0 (n = 1) M1 (n = 2)
Treatment intent = curative (n = 1) M0 (n = 0) M1 (n = 1)
Not documented (n = 1)
Treatment intent = palliative (n = 4) M0 (n = 1) M1 (n = 3)
FIGURE 1. Consort diagram regarding treatment intent for all patients before and after long-course chemoradiation (LCCRT).
of patients with a change in management). The use of PET scan and MRI restaging are highlighted by the current study, with PET-CT scan altering distant restaging more, whereas MRI restaging impacts local restaging. In addition, notable positive findings from the study included 7 patients where additional chemotherapy was administered before planned surgery, 8 patients who underwent more extensive surgical resections, and 5 additional patients in whom intraoperative radiation was offered after restaging. The study strengths include a standardized institutional protocol whereby all of the patients were restaged with all modalities at the same time interval (4 weeks), with prospective documentation of tumor board meetings to guide management. Our study is primarily limited by its single-center, retrospective design. In addition, given that the records were reviewed retrospectively, a few cases did lack clear management and operative details. The imaging was also reported by numerous radiologists and nuclear medicine technicians over the time period. However, we feel that the
conclusions represent a real-world experience of restaging as a routine component of the clinical management of rectal cancer. Another potential critique is the timing of the restaging. In line with the current literature, as an institution we have extended the interval from cessation of CRT to definitive surgery. A logical question would be whether we are restaging too early. Restaging was undertaken at 4 weeks post-LCCRT. Surgery was initially done at 8 weeks and eventually lengthened to 10 weeks in accordance with the literature. We agree that this leads to a lower accuracy rate compared with histological staging given the time interval. The department has addressed this, and timing of restaging is being reviewed. Restaging local disease remains a challenge, because there is variable accuracy. In our series, there was an accuracy of 43% for T stage and 56% to 60% for N stage as seen in Table 3. However, local staging does not change management in most cases, and with the exception of large tumors or those that threaten resection margins, management of local disease is often guided by initial staging. Conversely,
TABLE 5. Change in management plan after neoadjuvant LCCRT Variable
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
No Yes Missing
145 (90% (84%–94%)) 17 (10% (6%–16%)) 1
22 (65% (46%–80%)) 12 (35% (20%–54%)) 2
All patients (N = 199), n (OR (95% CI)) 167 (85% (79%–90%)) 29 (15% (10%–21%)) 3
LCCRT = long-course chemoradiation.
Copyright © The American Society of Colon & Rectal Surgeons, Inc. Unauthorized reproduction of this article is prohibited.
184
Schneider et al: Rectal Cancer Restaging After Chemoradiation
TABLE 6. Change in management plan after neoadjuvant LCCRT: responsible imaging modalities Variable PET No Yes Missing MRI No Yes Missing CT No Yes Missing
M0 (N = 163), n (OR (95% CI))
M1 (N = 36), n (OR (95% CI))
All patients (N = 199), n (OR (95% CI))
151 (93% (88%–97%)) 11 (7% (3%–12%)) 1
23 (68% (49%–83%)) 11 (32% (17%–51%)) 2
174 (89% (84%–93%)) 22 (11% (7%–16%)) 3
156 (96% (92%–99%)) 6 (4% (1%–8%)) 1
32 (94% (80%–99%)) 2 (6% (1%–20%)) 2
188 (96% (92%–98%)) 8 (4% (2%–8%)) 3
160 (99% (96%–100%)) 2 (1% (0%–4%)) 1
28 (82% (65%–93%)) 6 (18% (7%–35%)) 2
188 (96% (92%–98%)) 8 (4% (2%–8%)) 3
LCCRT = long-course chemoradiation; PET = positron emission tomography.
staging of distant disease is more accurate. Although we had a limited number of patients who had a metastasectomy to assess the accuracy of restaging, given that not all metastatic disease identified was resectable, the M-stage accuracy was
