Journal of Surgical Oncology 2014;109:652–658
Intraoperative Radiation Therapy for Locally Advanced Primary and Recurrent Colorectal Cancer: Ten-Year Institutional Experience JOHN R. HYNGSTROM, MD,1 CHING-WEI D. TZENG, MD,2 SAM BEDDAR, PhD,3 PRAJNAN DAS, MD, MPH,4 SUNIL KRISHNAN, MD,4 MARC E. DELCLOS, MD,4 CHRISTOPHER H. CRANE, MD,4 GEORGE J. CHANG, MD, MPH,5 Y. NANCY YOU, MD, MHSc,5 BARRY W. FEIG, MD,5 JOHN M. SKIBBER, MD,5 5 AND MIGUEL A. RODRIGUEZ-BIGAS, MD * 1
Section of Surgical Oncology, Department of Surgery, University of Utah, Salt Lake City, Utah Section of Surgical Oncology, Department of Surgery, University of Kentucky, Lexington, Kentucky 3 Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, Texas 4 Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 5 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 2
Background: We evaluated the role of intraoperative radiation therapy (IORT) during radical resection of locally advanced colorectal cancer (CRC). Methods: We retrospectively evaluated all patients with CRC treated with IORT at our institution from 2001 to 2010. IORT was delivered using high‐dose‐rate brachytherapy (median 12.5‐Gy). We analyzed factors associated with postoperative morbidity, local control (LC), and overall survival (OS). Results: One hundred patients were evaluated with 70% received IORT for recurrent tumors. R0 resection rate was 58%. Postoperative Grade 3 complications (33%) were independently associated with transfusions 3 units packed red blood cells (P ¼ 0.016) and body mass index (BMI) 35 (P ¼ 0.0499). Eighty‐two patients underwent external beam radiation therapy (EBRT) before IORT. Five‐year LC was 94%, for primary and 56%, for recurrent tumors, respectively (P ¼ 0.007). Microscopic positive (R1) margins were not associated with LC (P ¼ 0.316). BMI 30 (P ¼ 0.048) and post‐discharge complications (P ¼ 0.041) were independent risk factors for worse LC. Median post‐IORT OS was 67.7 (95% CI 51.1–84.3) months for all patients. Conclusion: For patients with primary or recurrent locally advanced CRC, treatment with radical surgery and IORT achieved excellent LC outcomes irrespective of microscopic margin status. IORT may be indicated for tumors suspected to have close or positive microscopic margins.
J. Surg. Oncol. 2014;109:652–658. ß 2014 Wiley Periodicals, Inc.
KEY WORDS: rectal; brachytherapy; margins; local recurrence; multimodality
INTRODUCTION Local control (LC) of advanced primary and locally recurrent colorectal cancer (CRC) is difficult to achieve with surgery alone, with historical relapse occurring in up to 50% of rectal cancer patients based on initial stage [1]. More recently, this rate is 11% among patients with rectal cancer undergoing total mesorectal excision (TME) alone [2]. Preoperative external beam radiation therapy (EBRT) improves LC by 50%, making it the standard of care for primary rectal tumors that are at least T3 [3]. Despite these improvements, local failures occur and can negatively affect quality of life and survival [4]. Although its use remains controversial, intraoperative radiation therapy (IORT) as an adjunct to multimodality therapy has been introduced during radical resection of locally advanced or recurrent CRC to improve LC for tumors at high risk for local failure. The theoretical advantage of IORT in addition to standard EBRT is its ability to increase the effective radiation dose to the tumor bed or margin at risk for local failure while excluding or limiting the dose to adjacent sensitive abdominal/pelvic structures [5]. IORT, delivered either as intraoperative external beam radiation or brachytherapy, has been used in a variety of malignancies, especially when primary tumor control has critical implications for patient morbidity and quality of life. Examples include selected patients with rectal cancer, sarcoma, and breast cancer [5]. It is less appropriate in patients with high likelihood of mortality from metastatic disease, such as pancreatic and gastric cancer [6,7]. IORT has been studied extensively for CRC, in which several large case series and one prospective trial have demonstrated 5‐year LC rates of 81–97% [8–16].
ß 2014 Wiley Periodicals, Inc.
Despite these achievements in improving LC, the widespread use of IORT has not been adopted due to concerns about increased postoperative morbidity and questionable additional benefit for patients who have received standard EBRT with proper TME. Rates of short‐term complications (bleeding, wound infections, fistulas, skin reactions) and long‐term morbidity (ureteral stenosis/obstruction, abscesses, chronic pain, sacral fractures, incontinence) have been consistently reported to be higher versus surgery or EBRT/surgery alone and seem to be associated with higher doses of IORT [8,9,11,13,14,17,18]. In the one randomized trial directly comparing outcomes of patients with rectal cancer receiving EBRT and TME with/without IORT, the addition of IORT showed no improvement in disease‐free or overall survival (OS) and was associated with higher perioperative morbidity [9]. At this study’s institution, IORT has been used only on an individualized basis for primary and recurrent locally advanced colorectal cancers in which the surgeon predicts that there is the
Conflict of interest: none. *Correspondence to: Miguel A. Rodriguez‐Bigas, MD, Department of Surgical Oncology, Unit 1484, The University of Texas MD Anderson Cancer Center, 1400 Pressler Street, Houston, TX 77030. Fax: þ713‐745‐ 5235. E‐mail: [email protected] Received 12 November 2013; Accepted 14 january 2014 DOI 10.1002/jso.23570 Published online 10 February 2014 in Wiley Online Library (wileyonlinelibrary.com).
IORT for Colorectal Cancer possibility of a positive microscopic margin on final pathology despite aggressive radical resection of pelvic contents. The hypothesis of this study was that IORT, when selectively used in a high‐risk cohort, could improve LC rates for patients who receive R1 resections. In this context, the primary purpose of this study was to evaluate the local control (LC) rates and the factors which affected LC in these highly selected patients. Secondary endpoints included evaluation of postoperative morbidity (and risk factors) and overall survival (OS).
METHODS
653
For primary cancers, most patients with rectal cancer routinely underwent standard (50.4 Gy) external beam radiation therapy (EBRT) before their index operation with IORT. There was a treatment break of 6–8 weeks before the index operation if EBRT was used preoperatively. Patients who did not receive standard‐dose EBRT for their primary tumors (e.g., primary sigmoid cancer) were referred for possible EBRT for the first time, with a treatment break of 6– 8 weeks before the index operation. These decisions were made in the setting of multidisciplinary discussions among surgical, radiation, and medical oncologists.
Patients, Staging, and Decision for IORT
Surgery and IORT
This study was approved by the University of Texas MD Anderson Cancer Center institutional review board. We retrospectively evaluated all consecutive patients with biopsy proven colorectal cancer treated with IORT using high‐dose‐rate (HDR) brachytherapy catheters at our institution from January 2001 to June 2010. All patients were radiographically staged with cross‐sectional imaging to determine the extent of locally advanced disease and the need for IORT. For both primary and recurrent cancer, patients preoperatively deemed at risk for possible microscopically positive margins were referred by the surgery team to the radiation oncology group. If R2 (gross residual tumor) resection was predicted, then both IORT and surgery were not offered. The “index” operation refers to the operation in which IORT was utilized along with radical resection.
Surgical procedures ranged from low anterior resections to total pelvic exenterations with en bloc resection of involved adjacent structures. All operations were performed with intent for R0, or negative microscopic margin, resection. R0 required at least a 2 mm margin by institutional definition. No R2 resections were planned, and none were performed. After the extirpative phase of the operation and before any anastomosis, flexible intraoperative brachytherapy catheters were placed along the margin at risk for microscopic positivity [19]. IORT was delivered using the Harrison–Anderson–Mick (HAM) applicator (Fig. 1) at a dose range of 10–15 Gy, prescribed at 1 cm from the source. Dosing was prescribed by the radiation oncologist based on the balance between the risks of local recurrence and toxicity to surrounding structures (e.g., having the ureter near the margin at risk). Surrounding structures, including the ureter and small bowel, were shielded with lead and wet laparotomy pads when appropriate. The HAM applicator was held in place with metal retractors or heavy surgical instruments to maximize the tumor bed surface apposition to the applicator. After completion of IORT, the HAM applicator, lead shields, and wet laparotomy pads were removed, and the operation was completed.
Pre‐IORT Radiation Therapy Patients who had previously received EBRT for primary tumors were usually given 39 Gy of re‐irradiation for their recurrence, while those who had never received EBRT received 50.4 Gy for recurrent tumors.
Fig. 1. A: High‐dose‐rate (HDR) radiation source and (B) Harrison–Anderson–Mick HAM applicator shown in perineal wound along margin at risk after abdominoperineal resection. Journal of Surgical Oncology
654
Hyngstrom et al. Postoperative Morbidity and Long‐Term Outcomes
Perioperative clinical factors were identified from the institutional colorectal cancer database when available and supplemented with the electronic medical record. Postoperative morbidity included 30‐day postoperative complications graded by Dindo–Clavien criteria [20]. Post‐discharge complications included any grade complication found within 90 days after index hospital discharge. Patients underwent cross‐ sectional imaging and/or endoscopic surveillance as indicated throughout their postoperative follow‐up. Long‐term oncologic outcomes included LC rates and OS rates, which were both measured from the date of surgery with IORT to the date of the respective event (recurrence or death) or last follow‐up if no event occurred before the last follow‐up.
Statistical Analysis Nonparametric univariate analyses were performed with chi‐squared test or Fisher’s exact test for categorical data and Mann–Whitney U‐test for continuous data to determine factors associated with postoperative morbidity. Log‐rank test was used to compare univariate factors associated with LC and OS. Significant univariate factors with P < 0.05 were entered into multivariate logistic regression models to find independent associations with major complications and local recurrence. Statistical significance was defined as two‐sided P < 0.05. Multivariate modeling was reported with odds ratio (OR) and 95% confidence interval (CI). Life table analyses were used to calculate actuarial long‐ term LC and OS. All analyses were performed using IBM SPSS Version 19 software (IBM, Armonk, NY).
RESULTS Patients and Tumors Using the multidisciplinary colorectal database, 100 consecutive patients with colorectal cancer (87 rectal, 13 colon) were identified with median age of 60 years (range 20–80) and median follow‐up 34.0 months (range 0.7–112.6). Fifty‐nine percent were male and 76% were Caucasian. Preoperative patient demographics and tumor characteristics are detailed in Table I.
Surgery and IORT R0 resection rate was 54% (R1 46%, R2 0%). There was no difference in R0 resection rate among patients with primary versus recurrent tumors (Table I, P ¼ 0.220). Seventy patients (70%) underwent surgery with IORT for recurrent cancer. IORT was delivered at a median dose of 12.5 Gy (range 10–15). Median interval between the primary tumor resection and the index operation with IORT for recurrence was 32.8 months (range 4.7–246.7). Seventy‐five patients had multivisceral en bloc resections of other structures (median n ¼ 2; range 1–7). Eighty‐two patients (82%) underwent EBRT before the index operation with IORT, and 38 of these 82 (46.3%) patients had received previous EBRT for their primary tumors. Of the 18 patients who did not receive EBRT before the operation with IORT, 9 (50%) had received previous EBRT. EBRT for primary tumors was delivered at a median dose of 50.4 Gy (range 39–59) in 28 fractions (range 25–30). EBRT for recurrent tumors was delivered at a median dose of 39 Gy (range 30–50.4) in 26 fractions (range 20–36).
Morbidity of Surgery With IORT The median operative time was 589 min (range 123–1,075) with median estimated blood loss (EBL) 1,100 c3 (range 200–7,550). There Journal of Surgical Oncology
TABLE I. Patients, Tumor Characteristics, and Operations Clinical characteristic Total patients Median age, years Gender, male Median body mass index ECOG performance status Albumin, g/dl Hemoglobin 10 Gy Positive margin (92% at 5 years) for patients receiving EBRT and TME alone, with no statistical difference when adding IORT. However, unlike the current study, the French study very few locally advanced patients, despite stating that all patients were “locally advanced” T3/T4, with only 7% of patients having T4 lesions, 3% poor differentiation, and none with recurrent tumors (versus 70% recurrent tumors in the current study) [9]. Thus it should not be a surprise that adding IORT to standard‐dose EBRT and TME is unnecessary in primary T3 rectal cancers. Our cohort represents a group of patients at higher risk of local recurrence. A unique and unexpected finding of our study was that LC and OS was no different in patients with either margin‐negative (R0) or microscopically margin‐positive (R1) resections with the use of IORT within the context of modern multimodality therapy. This finding contrasts sharply with the two largest American series of patients with locally advanced (T4) or recurrent CRC with long‐term follow‐ up [12,21]. The Massachusetts General Hospital (MGH) experience was remarkable for reporting 89% and 68% 5‐year LC for R0/R1 resections, respectively [12]. Our patient population differs in that the majority of our patients had recurrent cancer (as opposed to none in the MGH study). In addition, treatment eras were different. Our patients were treated after 2001 using HDR brachytherapy while the MGH study included only patients from 1996, treated with external IORT using electron beams (IOERT). Additionally, with the more contemporary study period (the most recent decade), the current study’s patients were treated with modern chemotherapy including oxaliplatin, which would not have been available for the MGH patients. The Mayo Clinic group published long‐term follow‐up of over 600 patients with locally recurrent CRC treated from 1981 to 2008, with 72% and 68% LC for R0 and R1 resection, respectively, and an OS difference favoring R0 resection [21]. Our more contemporary series differs as we included only patients after 2001, all of whom received brachytherapy using a HAM applicator as opposed to the Mayo study which used only IOERT. Also, 16% of patients with R2 resections were included in the Mayo analysis, whereas our strategy did not include any R2 resections. In the setting of modern systemic chemotherapy and its positive effect upon OS, the survival advantage of a locoregional treatment such as IORT seen in previously reported series from the 1980s and 1990s was not seen in our
656
Hyngstrom et al. TABLE III. Univariate and Multivariate Factors Associated With Post‐ IORT Local Control Clinical characteristic
Grade poorly differentiated EBRT before this operation Recurrent tumor BMI 30 IORT 10 Gray En bloc organs 3 Positive margins (
Intraoperative Radiation Therapy for Locally Advanced Primary and Recurrent Colorectal Cancer: Ten-Year Institutional Experience JOHN R. HYNGSTROM, MD,1 CHING-WEI D. TZENG, MD,2 SAM BEDDAR, PhD,3 PRAJNAN DAS, MD, MPH,4 SUNIL KRISHNAN, MD,4 MARC E. DELCLOS, MD,4 CHRISTOPHER H. CRANE, MD,4 GEORGE J. CHANG, MD, MPH,5 Y. NANCY YOU, MD, MHSc,5 BARRY W. FEIG, MD,5 JOHN M. SKIBBER, MD,5 5 AND MIGUEL A. RODRIGUEZ-BIGAS, MD * 1
Section of Surgical Oncology, Department of Surgery, University of Utah, Salt Lake City, Utah Section of Surgical Oncology, Department of Surgery, University of Kentucky, Lexington, Kentucky 3 Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, Texas 4 Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 5 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas 2
Background: We evaluated the role of intraoperative radiation therapy (IORT) during radical resection of locally advanced colorectal cancer (CRC). Methods: We retrospectively evaluated all patients with CRC treated with IORT at our institution from 2001 to 2010. IORT was delivered using high‐dose‐rate brachytherapy (median 12.5‐Gy). We analyzed factors associated with postoperative morbidity, local control (LC), and overall survival (OS). Results: One hundred patients were evaluated with 70% received IORT for recurrent tumors. R0 resection rate was 58%. Postoperative Grade 3 complications (33%) were independently associated with transfusions 3 units packed red blood cells (P ¼ 0.016) and body mass index (BMI) 35 (P ¼ 0.0499). Eighty‐two patients underwent external beam radiation therapy (EBRT) before IORT. Five‐year LC was 94%, for primary and 56%, for recurrent tumors, respectively (P ¼ 0.007). Microscopic positive (R1) margins were not associated with LC (P ¼ 0.316). BMI 30 (P ¼ 0.048) and post‐discharge complications (P ¼ 0.041) were independent risk factors for worse LC. Median post‐IORT OS was 67.7 (95% CI 51.1–84.3) months for all patients. Conclusion: For patients with primary or recurrent locally advanced CRC, treatment with radical surgery and IORT achieved excellent LC outcomes irrespective of microscopic margin status. IORT may be indicated for tumors suspected to have close or positive microscopic margins.
J. Surg. Oncol. 2014;109:652–658. ß 2014 Wiley Periodicals, Inc.
KEY WORDS: rectal; brachytherapy; margins; local recurrence; multimodality
INTRODUCTION Local control (LC) of advanced primary and locally recurrent colorectal cancer (CRC) is difficult to achieve with surgery alone, with historical relapse occurring in up to 50% of rectal cancer patients based on initial stage [1]. More recently, this rate is 11% among patients with rectal cancer undergoing total mesorectal excision (TME) alone [2]. Preoperative external beam radiation therapy (EBRT) improves LC by 50%, making it the standard of care for primary rectal tumors that are at least T3 [3]. Despite these improvements, local failures occur and can negatively affect quality of life and survival [4]. Although its use remains controversial, intraoperative radiation therapy (IORT) as an adjunct to multimodality therapy has been introduced during radical resection of locally advanced or recurrent CRC to improve LC for tumors at high risk for local failure. The theoretical advantage of IORT in addition to standard EBRT is its ability to increase the effective radiation dose to the tumor bed or margin at risk for local failure while excluding or limiting the dose to adjacent sensitive abdominal/pelvic structures [5]. IORT, delivered either as intraoperative external beam radiation or brachytherapy, has been used in a variety of malignancies, especially when primary tumor control has critical implications for patient morbidity and quality of life. Examples include selected patients with rectal cancer, sarcoma, and breast cancer [5]. It is less appropriate in patients with high likelihood of mortality from metastatic disease, such as pancreatic and gastric cancer [6,7]. IORT has been studied extensively for CRC, in which several large case series and one prospective trial have demonstrated 5‐year LC rates of 81–97% [8–16].
ß 2014 Wiley Periodicals, Inc.
Despite these achievements in improving LC, the widespread use of IORT has not been adopted due to concerns about increased postoperative morbidity and questionable additional benefit for patients who have received standard EBRT with proper TME. Rates of short‐term complications (bleeding, wound infections, fistulas, skin reactions) and long‐term morbidity (ureteral stenosis/obstruction, abscesses, chronic pain, sacral fractures, incontinence) have been consistently reported to be higher versus surgery or EBRT/surgery alone and seem to be associated with higher doses of IORT [8,9,11,13,14,17,18]. In the one randomized trial directly comparing outcomes of patients with rectal cancer receiving EBRT and TME with/without IORT, the addition of IORT showed no improvement in disease‐free or overall survival (OS) and was associated with higher perioperative morbidity [9]. At this study’s institution, IORT has been used only on an individualized basis for primary and recurrent locally advanced colorectal cancers in which the surgeon predicts that there is the
Conflict of interest: none. *Correspondence to: Miguel A. Rodriguez‐Bigas, MD, Department of Surgical Oncology, Unit 1484, The University of Texas MD Anderson Cancer Center, 1400 Pressler Street, Houston, TX 77030. Fax: þ713‐745‐ 5235. E‐mail: [email protected] Received 12 November 2013; Accepted 14 january 2014 DOI 10.1002/jso.23570 Published online 10 February 2014 in Wiley Online Library (wileyonlinelibrary.com).
IORT for Colorectal Cancer possibility of a positive microscopic margin on final pathology despite aggressive radical resection of pelvic contents. The hypothesis of this study was that IORT, when selectively used in a high‐risk cohort, could improve LC rates for patients who receive R1 resections. In this context, the primary purpose of this study was to evaluate the local control (LC) rates and the factors which affected LC in these highly selected patients. Secondary endpoints included evaluation of postoperative morbidity (and risk factors) and overall survival (OS).
METHODS
653
For primary cancers, most patients with rectal cancer routinely underwent standard (50.4 Gy) external beam radiation therapy (EBRT) before their index operation with IORT. There was a treatment break of 6–8 weeks before the index operation if EBRT was used preoperatively. Patients who did not receive standard‐dose EBRT for their primary tumors (e.g., primary sigmoid cancer) were referred for possible EBRT for the first time, with a treatment break of 6– 8 weeks before the index operation. These decisions were made in the setting of multidisciplinary discussions among surgical, radiation, and medical oncologists.
Patients, Staging, and Decision for IORT
Surgery and IORT
This study was approved by the University of Texas MD Anderson Cancer Center institutional review board. We retrospectively evaluated all consecutive patients with biopsy proven colorectal cancer treated with IORT using high‐dose‐rate (HDR) brachytherapy catheters at our institution from January 2001 to June 2010. All patients were radiographically staged with cross‐sectional imaging to determine the extent of locally advanced disease and the need for IORT. For both primary and recurrent cancer, patients preoperatively deemed at risk for possible microscopically positive margins were referred by the surgery team to the radiation oncology group. If R2 (gross residual tumor) resection was predicted, then both IORT and surgery were not offered. The “index” operation refers to the operation in which IORT was utilized along with radical resection.
Surgical procedures ranged from low anterior resections to total pelvic exenterations with en bloc resection of involved adjacent structures. All operations were performed with intent for R0, or negative microscopic margin, resection. R0 required at least a 2 mm margin by institutional definition. No R2 resections were planned, and none were performed. After the extirpative phase of the operation and before any anastomosis, flexible intraoperative brachytherapy catheters were placed along the margin at risk for microscopic positivity [19]. IORT was delivered using the Harrison–Anderson–Mick (HAM) applicator (Fig. 1) at a dose range of 10–15 Gy, prescribed at 1 cm from the source. Dosing was prescribed by the radiation oncologist based on the balance between the risks of local recurrence and toxicity to surrounding structures (e.g., having the ureter near the margin at risk). Surrounding structures, including the ureter and small bowel, were shielded with lead and wet laparotomy pads when appropriate. The HAM applicator was held in place with metal retractors or heavy surgical instruments to maximize the tumor bed surface apposition to the applicator. After completion of IORT, the HAM applicator, lead shields, and wet laparotomy pads were removed, and the operation was completed.
Pre‐IORT Radiation Therapy Patients who had previously received EBRT for primary tumors were usually given 39 Gy of re‐irradiation for their recurrence, while those who had never received EBRT received 50.4 Gy for recurrent tumors.
Fig. 1. A: High‐dose‐rate (HDR) radiation source and (B) Harrison–Anderson–Mick HAM applicator shown in perineal wound along margin at risk after abdominoperineal resection. Journal of Surgical Oncology
654
Hyngstrom et al. Postoperative Morbidity and Long‐Term Outcomes
Perioperative clinical factors were identified from the institutional colorectal cancer database when available and supplemented with the electronic medical record. Postoperative morbidity included 30‐day postoperative complications graded by Dindo–Clavien criteria [20]. Post‐discharge complications included any grade complication found within 90 days after index hospital discharge. Patients underwent cross‐ sectional imaging and/or endoscopic surveillance as indicated throughout their postoperative follow‐up. Long‐term oncologic outcomes included LC rates and OS rates, which were both measured from the date of surgery with IORT to the date of the respective event (recurrence or death) or last follow‐up if no event occurred before the last follow‐up.
Statistical Analysis Nonparametric univariate analyses were performed with chi‐squared test or Fisher’s exact test for categorical data and Mann–Whitney U‐test for continuous data to determine factors associated with postoperative morbidity. Log‐rank test was used to compare univariate factors associated with LC and OS. Significant univariate factors with P < 0.05 were entered into multivariate logistic regression models to find independent associations with major complications and local recurrence. Statistical significance was defined as two‐sided P < 0.05. Multivariate modeling was reported with odds ratio (OR) and 95% confidence interval (CI). Life table analyses were used to calculate actuarial long‐ term LC and OS. All analyses were performed using IBM SPSS Version 19 software (IBM, Armonk, NY).
RESULTS Patients and Tumors Using the multidisciplinary colorectal database, 100 consecutive patients with colorectal cancer (87 rectal, 13 colon) were identified with median age of 60 years (range 20–80) and median follow‐up 34.0 months (range 0.7–112.6). Fifty‐nine percent were male and 76% were Caucasian. Preoperative patient demographics and tumor characteristics are detailed in Table I.
Surgery and IORT R0 resection rate was 54% (R1 46%, R2 0%). There was no difference in R0 resection rate among patients with primary versus recurrent tumors (Table I, P ¼ 0.220). Seventy patients (70%) underwent surgery with IORT for recurrent cancer. IORT was delivered at a median dose of 12.5 Gy (range 10–15). Median interval between the primary tumor resection and the index operation with IORT for recurrence was 32.8 months (range 4.7–246.7). Seventy‐five patients had multivisceral en bloc resections of other structures (median n ¼ 2; range 1–7). Eighty‐two patients (82%) underwent EBRT before the index operation with IORT, and 38 of these 82 (46.3%) patients had received previous EBRT for their primary tumors. Of the 18 patients who did not receive EBRT before the operation with IORT, 9 (50%) had received previous EBRT. EBRT for primary tumors was delivered at a median dose of 50.4 Gy (range 39–59) in 28 fractions (range 25–30). EBRT for recurrent tumors was delivered at a median dose of 39 Gy (range 30–50.4) in 26 fractions (range 20–36).
Morbidity of Surgery With IORT The median operative time was 589 min (range 123–1,075) with median estimated blood loss (EBL) 1,100 c3 (range 200–7,550). There Journal of Surgical Oncology
TABLE I. Patients, Tumor Characteristics, and Operations Clinical characteristic Total patients Median age, years Gender, male Median body mass index ECOG performance status Albumin, g/dl Hemoglobin 10 Gy Positive margin (92% at 5 years) for patients receiving EBRT and TME alone, with no statistical difference when adding IORT. However, unlike the current study, the French study very few locally advanced patients, despite stating that all patients were “locally advanced” T3/T4, with only 7% of patients having T4 lesions, 3% poor differentiation, and none with recurrent tumors (versus 70% recurrent tumors in the current study) [9]. Thus it should not be a surprise that adding IORT to standard‐dose EBRT and TME is unnecessary in primary T3 rectal cancers. Our cohort represents a group of patients at higher risk of local recurrence. A unique and unexpected finding of our study was that LC and OS was no different in patients with either margin‐negative (R0) or microscopically margin‐positive (R1) resections with the use of IORT within the context of modern multimodality therapy. This finding contrasts sharply with the two largest American series of patients with locally advanced (T4) or recurrent CRC with long‐term follow‐ up [12,21]. The Massachusetts General Hospital (MGH) experience was remarkable for reporting 89% and 68% 5‐year LC for R0/R1 resections, respectively [12]. Our patient population differs in that the majority of our patients had recurrent cancer (as opposed to none in the MGH study). In addition, treatment eras were different. Our patients were treated after 2001 using HDR brachytherapy while the MGH study included only patients from 1996, treated with external IORT using electron beams (IOERT). Additionally, with the more contemporary study period (the most recent decade), the current study’s patients were treated with modern chemotherapy including oxaliplatin, which would not have been available for the MGH patients. The Mayo Clinic group published long‐term follow‐up of over 600 patients with locally recurrent CRC treated from 1981 to 2008, with 72% and 68% LC for R0 and R1 resection, respectively, and an OS difference favoring R0 resection [21]. Our more contemporary series differs as we included only patients after 2001, all of whom received brachytherapy using a HAM applicator as opposed to the Mayo study which used only IOERT. Also, 16% of patients with R2 resections were included in the Mayo analysis, whereas our strategy did not include any R2 resections. In the setting of modern systemic chemotherapy and its positive effect upon OS, the survival advantage of a locoregional treatment such as IORT seen in previously reported series from the 1980s and 1990s was not seen in our
656
Hyngstrom et al. TABLE III. Univariate and Multivariate Factors Associated With Post‐ IORT Local Control Clinical characteristic
Grade poorly differentiated EBRT before this operation Recurrent tumor BMI 30 IORT 10 Gray En bloc organs 3 Positive margins (
