Potential Benefits of Integration of Therapies With Chemoradiation in Rectal Cancer

Abstract

Improved pelvic control with reduced toxicity and enhanced sphincter preservation has been demonstrated with neoadjuvant chemoradiation compared with postoperative adjuvant chemoradiation in patients with stage II and III rectal cancer. However, analyses from many trials of adjuvant chemoradiation indicate that patients with T3 node-positive and T4 tumors are at high risk of pelvic recurrence even with use of chemoradiation. This limitation could be addressed with treatment intensification strategies, such as increasing the radiotherapy dose using altered fractionation, or incorporation of novel cytotoxic and targeted chemotherapeutic agents. In addition, preliminary evidence suggests that selected patients with clinically staged T2 or T3 node-negative tumors may be candidates for trials evaluating organ-preserving strategies after chemoradiation, thus eliminating the morbidity of radical surgery. These efforts could be enhanced with the availability of more effective chemoradiation regimens. This paper will discuss incorporation of molecular targeted therapy with chemoradiation regimens in the context of current standards, limitations, and new concepts in the combined modality therapy of locally advanced rectal cancer.

Based on randomized comparisons of neoadjuvant vs. postoperative chemoradiation, neoadjuvant chemoradiation has become part of the standard of care for patients with locally advanced rectal cancer (T3/T4 or node-positive disease).¹ Trials that established the standard role of adjuvant chemoradiation in rectal cancer included all patients with locally advanced rectal cancer without stratification based on T or N stage.^2,3 A more recent pooled analysis of phase III rectal cancer trials suggests that patients with T3/N+ and T4 disease have particularly high rates of pelvic recurrence, even when treated with chemoradiation.⁴ These high-risk features also apply in the neoadjuvant setting, although our ability to accurately stage patients clinically has limitations. Neoadjuvant chemoradiation leads to improved sphincter preservation and pelvic control. Enhancement of chemoradiation through selective radiosensitization could benefit high-risk patients through reduction of pelvic recurrence and possibly improve overall survival through treatment intensification. Another possible benefit of enhancement of local treatment effect is the development of organ preservation strategies in locally advanced rectal cancer patients, although this is controversial and the multi-institutional feasibility of this concept has not been demonstrated. Incorporation of molecular targeted therapy with established chemoradiation regimens may effectively address these objectives.

CHEMORADIATION: CURRENT STATUS

The Evidence for Adjuvant Chemoradiation

The role of adjuvant therapy in rectal cancer has been evaluated in several randomized clinical trials, the first of which was conducted by the Gastrointestinal Tumor Study Group (GITSG). Patients with resected stage B2 or C (American Joint Committee on Cancer [AJCC] stages II and III) rectal cancer were randomly assigned to one of four groups: no adjuvant therapy, adjuvant radiotherapy (40–48 Gy), adjuvant chemotherapy (bolus 5-fluorouracil [5-FU] and semustine), or concurrent radiotherapy and chemotherapy.⁵ Patients treated with concurrent radiotherapy and chemotherapy had improved recurrence-free survival compared with the surgery-only group. Patients who received only radiotherapy had a lower rate of local failure, and those who received only chemotherapy had a lower rate of distant disease than those who received no adjuvant therapy. Subsequent analysis showed that adjuvant chemoradiation increased overall survival rates (45% with surgery only and 58% with adjuvant chemoradiation).⁶ The Mayo/North Central Cancer Treatment Group (NCCTG) trial subsequently randomized a similar population of patients to receive postoperative radiotherapy (45–50.4 Gy in 25–28 fractions) alone, or radiotherapy and concurrent bolus 5-FU, with a cycle of 5-FU and semustine before and after chemoradiation.² Adjuvant chemoradiation yielded significantly lower rates of both local recurrence and distant metastasis compared with radiation therapy alone. Moreover, adjuvant chemoradiation reduced the overall death rate by 29% relative to radiotherapy only.

In the National Surgical Adjuvant Breast and Bowel Project (NSABP) R-01 trial, patients with resected Dukes B or C rectal cancer were randomly assigned to one of three arms: no adjuvant therapy, postoperative radiotherapy (46–47 Gy in 26–27 fractions), or postoperative chemotherapy (semustine, vincristine, and 5-FU).⁷ Compared with no adjuvant therapy, chemotherapy yielded higher disease-free and overall survival rates, but the effect was restricted to males; and radiotherapy decreased the locoregional recurrence rate from 25% to 16%, but did not improve disease-free or overall survival. In the subsequent NSABP R-02 trial, patients were randomized to receive postoperative chemotherapy with or without radiation.⁸ The addition of radiation decreased the locoregional relapse rate from 13% to 8%, but did not increase disease-free or overall survival.

In summary, these phase III studies showed that (1) adjuvant chemoradiation increases overall survival compared with no adjuvant treatment, (2) adjuvant chemoradiation improves overall survival compared with radiotherapy alone, and (3) adjuvant chemoradiation provides better locoregional control than chemotherapy alone. These randomized studies have clearly demonstrated the benefits of adjuvant chemoradiation in patients with resected stage T3/T4 and/or node-positive rectal cancer.

Risk Stratification: Pooled Analysis

The trials that established the benefit of adjuvant therapy in locally advanced rectal cancer included unstratified populations of patients with T3/T4 tumors or with positive lymph nodes. These trials also included a variety of treatment arms. With publication of the NSABP R-02 trial, it became apparent that all patients did not derive a survival benefit from the addition of chemoradiation to adjuvant chemotherapy. Rather, only patients with tumors arising in the low pelvis requiring an abdominoperineal resection (P = .07) and those less than 60 years old (P = .007) appeared to have improved overall survival on multivariate analysis.⁸ Subsequently, data from the NSABP R-02 and four other postoperative adjuvant therapy trials were combined to identify patient subgroups with different recurrence risks, based on T and N stage and treatment modality.⁴ This important “pooled analysis” has identified patient subgroups with high, moderately high, and intermediate risk of pelvic recurrence, and the risk level appears to correlate with potential benefits of chemoradiation, depending on the T and N stage. In particular, patients with T3 node-positive or T4 tumors who receive adjuvant chemoradiation appear to have improved pelvic control and overall survival, although pelvic recurrence rates are still high (10%–33%). On the other hand, patients with T3N0 and T2 node-positive disease do not all have an overall survival benefit, probably because the a priori risk of pelvic recurrence is lower. Local relapse rate was 5% for stage T1-2N1 and 11% for stage T3N0 patients treated with surgery and chemotherapy, but no radiotherapy.⁴ It is important to emphasize that these are heterogeneous groups of patients, and other factors such as radial margin status,⁹ quality of the surgery,^10,11 and extent of lymph-node sampling¹² affect recurrence risk. In addition, clinical factors relating to the anatomy of the pelvis, such as distance of tumor from the anal verge and tumor extension beyond the rectal wall, may also influence outcomes.¹³

Over the decades these randomized trials were conducted, advances in imaging have affected patient selection, and other technologic and supportive care approaches have improved. For these reasons, as well as differences in the rigor with which pelvic recurrence data were collected in the trials, the pooled analysis can only be considered hypothesis-generating. Prospective data from trials designed exclusively to test questions from the subgroups identified in the analysis are not available. Therefore, while the pooled analysis has not changed the standard of care, it should affect the design of future trials. In particular, trials addressing the use of preoperative or postoperative novel radiosensitizers should include the highest-risk patients to increase the chance of detecting a reduction in pelvic recurrence rate.

Patients at Low Risk for Local Recurrence

Results of a retrospective study from Massachusetts General Hospital (MGH) showed a 95% local control rate in patients with T3N0 tumors and favorable histologic features (well to moderately differentiated, < 2 mm invasion into perirectal fat, and no lymphovascular invasion) who underwent surgery without adjuvant chemotherapy or radiation.¹³ Furthermore, the Dutch Colorectal trial showed that among patients undergoing surgery alone, those with high rectal tumors (> 10 cm from the anal verge) and those with wide negative radial margins (> 2 mm) had a low risk of locoregional recurrence.^14,15 The number of lymph nodes in the resected specimen has also been associated with relapse risk.¹⁶ Thus, a selected subgroup of patients with T3N0 tumors with favorable pathologic features (well to moderate differentiation, minimal perirectal invasion, no lymphovascular invasion, > 10 cm from anal verge, widely negative radial margins, and adequate lymph-node dissection) may have a low risk of local relapse with surgery and no radiotherapy. Prospective trials are needed to evaluate whether radiotherapy may be safely omitted for selected patients with T3N0 or T2N1 disease.

Neoadjuvant vs. Postoperative Adjuvant Chemoradiation

First principles of radiotherapy have predicted for some time, based on preclinical data and limited clinical data, that chemoradiation should be better tolerated and more efficacious when administered preoperatively rather than postoperatively in rectal cancer patients. In the postoperative setting, the small bowel is often fixed in the pelvis, increasing the bowel volume that is irradiated; the tumor bed is hypoxic, limiting therapeutic efficacy; and the vasculature is interrupted, compromising chemotherapy delivery.

The German Chirurgische Arbeitsgemeinschaft Onkologie/Arbeitsgemeinschaft Radioonkologie/Arbeitsgemeinschaft Internistische Onkologie (CAO/ARO/AIO) trial comparing preoperative and postoperative chemoradiation has confirmed all of these hypotheses.¹ In this study, more than 800 patients were staged with computed tomography (CT) scan and endoscopic ultrasound, and those with clinical stage T3 or T4 tumors were randomly assigned to receive preoperative vs. postoperative chemoradiation (50.4 Gy in 28 fractions with concurrent high-dose infusional 5-FU [1,000mg/m²/day on days 1–5 and 21–25]). All patients underwent total mesorectal excision 6 weeks after completing chemoradiation, while some patients in the postoperative arm received an additional 5.4-Gy boost to the tumor bed. Patients in both arms were scheduled to receive four cycles of bolus 5-FU (500 mg/m²/day, five times weekly, every 4 weeks), either after surgery (preoperative arm) or after chemoradiation (postoperative arm). Results showed a lower pelvic recurrence rate in the preoperative chemoradiation arm (6% vs. 13% postoperative, respectively, P = .0006), even though approximately 25% more patients than in the postoperative arm had low-lying tumors. In addition, patients receiving preoperative therapy had lower rates of grade 3 or 4 acute (27% vs. 40%, P = .001) and late (14% vs. 24%, P = .01) toxicities. The most striking toxicity difference between the arms was the incidence of anastomotic strictures (4% vs. 12%, P < .003). Overall survival was unchanged. In patients who, based on clinical evaluation, were felt to require abdominoperineal resection, chemoradiation also led to increased sphincter preservation rates (39% vs. 19%, P = .004) and less evidence of anastomotic stricture.¹

The reasons for poorer local control in the postoperative chemoradiation group could have been due to hypoxia in the tumor bed limiting efficacy, as predicted by preclinical modeling and numerous clinical studies. Moreover, poor compliance in the postoperative arm due to toxicity may have contributed to the difference. Only 54%of patients in the postoperative arm received the full radiation dose and 50% received full-dose chemotherapy, compared with 92% and 89%, respectively, in the preoperative arm (P < .001). Regardless of the reason for success in the neoadjuvant chemoradiotherapy arm, this trial has been widely interpreted as having defined a new standard of care in the adjuvant treatment of rectal cancer.

One disadvantage associated with preoperative therapy is the potential overtreatment of patients whose clinical stage is initially judged higher than that determined pathologically at the time of surgery. For example, 18% of patients randomized to the postoperative arm were spared adjuvant therapy because they were found to have stage I disease at the time of surgery, and another 10% of patients did not receive adjuvant treatment due to detection of metastatic disease or postoperative complications/death.¹ This suggests that 15% to 20% of patients receiving preoperative chemoradiation may be over-treated.

CONCURRENT CYTOTOXIC CHEMOTHERAPY CONSIDERATIONS

Fluorouracil-Based Concurrent Chemoradiation

Until recently, the only phase III data supporting neoadjuvant use of concurrent chemotherapy with radiotherapy were extrapolated from trials evaluating postoperative therapy,^2,5 which showed improved outcome when 5-FU–based chemotherapy was added to radiation. Two recently reported randomized phase III trials evaluated the addition of chemotherapy to neoadjuvant standard fractionation radiotherapy alone. In the European Organization for Research and Treatment of Cancer (EORTC) 22921 study, 1,011 patients with resectable T3 or T4 rectal cancer were randomly assigned to receive preoperative radiotherapy with or without concurrent 5-FU and leucovorin.¹⁷ A second randomization evaluated adjuvant chemotherapy. Patients treated with preoperative chemoradiation had a higher pathologic complete response rate (14% vs. 5%), and local relapse rates were lower in chemotherapy-containing arms (8%–10%) than in the radiotherapy-alone arm (17%). Similar results were shown in the French phase III Fondation Française de Cancérologie Digestive (FFCD) 9203 trial,¹⁸ wherein 762 patients with resectable T3 or T4 rectal cancer received preoperative radiotherapy or preoperative chemoradiation. The addition of chemotherapy significantly increased the pathologic complete response rate from 4% to 12%, and decreased the local failure rate from 17% to 8% relative to the radiotherapy-alone group. Neither study showed improved event-free or overall survival with the addition of chemotherapy to radiotherapy. Nevertheless, both trials have clearly demonstrated that the addition of chemotherapy to standard fractionation radiotherapy in the preoperative setting increases local control and pathologic complete response rates.

The initial studies of concurrent chemoradiation all used bolus 5-FU alone or combined with other drugs. The Gastrointestinal Intergroup trial compared protracted venous infusional 5-FU to bolus 5-FU in patients with resected stage II or III rectal cancer.¹⁹ Patients were randomly assigned to receive bolus 5-FU (500 mg/m²) for 3 consecutive days during weeks 1 and 5 of radiation therapy, or protracted venous infusional 5-FU (225 mg/m²/day), 7 days every week during radiation therapy. No significant difference in local recurrence rate was noted for the two study arms, whereas protracted infusional 5-FU (vs. bolus 5-FU) significantly increased relapsefree (63% vs. 53%) and overall (70% vs. 60%) survival rates. Based on this trial, infusional 5-FU has been accepted as standard chemotherapy for concurrent chemoradiation in rectal cancer patients.

Subsequent trials have evaluated the biochemical modulation of 5-FU and incorporation of infusional 5-FU before and after chemoradiation. The Intergroup 0114 trial investigated whether the addition of leucovorin and/or levamisole would enhance efficacy of combined postoperative radiation and bolus 5-FU, and results showed no significant advantage over that achieved with bolus 5-FU alone.^20,21 The Intergroup 0144 trial compared three 5-FU regimens (bolus 5-FU, protracted infusional 5-FU, and 5-FU with leucovorin and levamisole), which were administered both before and after chemoradiation. Final results showed no significant differences in relapse-free or overall survival, and pelvic recurrence rates were less than 10% in all three arms.²² From this trial, it appears that use of biochemically modulated 5-FU and protracted infusional 5-FU result in similar outcomes.

Capecitabine, an oral agent that pharmacologically mimics protracted venous infusional 5-FU, has been evaluated in recent studies. In a phase III comparison of adju vant capecitabine vs. bolus 5-FU/leucovorin in stage III colon cancer, capecitabine yielded similar disease-free survival, with a trend toward greater disease-free and overall survival, as well as lower rates of gastrointestinal toxicity, neutropenia, and stomatitis.²³ In several phase I and II investigations of concurrent preoperative chemoradiation with capecitabine,^24–31 this agent resulted in pathologic complete response rates of 12% to 24% and an acceptable toxicity profile. A matched-pair comparison showed no significant differences in grade 3 or 4 toxicities, pathologic response, sphincter preservation, and relapse rates in patients who received preoperative radiotherapy with either concurrent capecitabine or concurrent protracted infusional 5-FU.³⁰ The dose-limiting toxicity of capecitabine is hand-foot syndrome, and this agent appears to cause fewer gastrointestinal and hematologic toxicities than intravenous, bolus, or infusional 5-FU. The recommended capecitabine dose is 825 mg/m² twice daily, including weekends, or 900 mg/m² twice daily on Monday through Friday continuously throughout the course of radiation. Thus, capecitabine appears to serve as an acceptable alternative to infusional 5-FU for preoperative chemoradiation in rectal cancer patients. The NSABP R-04 trial is a randomized comparison of preoperative radiotherapy with either concurrent capecitabine or concurrent protracted infusional 5-FU.

Irinotecan vs. Oxaliplatin as Radiosensitizers in Rectal Cancer

Oxaliplatin, a third-generation platinum analogue with radiosensitizing properties, was approved by the US Food and Drug Administration (FDA) for adjuvant treatment of colon cancers based on results of the MOSAIC (Multicenter International Study of Oxaliplatin/5-Fluorouracil/Leucovorin in the Adjuvant Treatment of Colon Cancer) trial.³² The potential for concurrent chemotherapy to have synergistic or additive activity on the primary tumor supports the study of oxaliplatin as a radiosensitizer. In phase I and II studies, oxaliplatin was well tolerated in combination with radiation and either capecitabine or infusional 5-FU in rectal cancer patients.^33,34 The addition of oxaliplatin to chemoradiation regimens could potentially benefit patients at high risk for relapse. The ongoing NSABP R-04 trial used a 2×2 factorial design, wherein patients were randomized to receive oxaliplatin or no oxaliplatin, and either infusional 5-FU or capecitabine, concurrently with radiotherapy. The primary trial end point is pelvic tumor control. Another randomized trial in France is comparing preoperative chemoradiation with capecitabine alone, or combined with oxaliplatin. Finally, the Radiation Therapy Oncology Group (RTOG) 0247 trial is a randomized phase II comparison of irinotecan vs. oxaliplatin in patients receiving capecitabine-based chemoradiation. These trials will provide more definitive evidence on the roles of capecitabine and oxaliplatin in preoperative therapy for rectal cancer.

In contrast, although several years ago it was rational to study irinotecan as a radiosensitizer because it was the first drug to show a median survival advantage in the metastatic setting as a single agent and combined with 5-FU, and it has radiosensitizing properties in the lab,³⁵ the subsequent failure of irinotecan and success of oxaliplatin³² and capecitabine³⁶ as adjuvant therapy for colon cancer have diminished the relative appeal of irinotecan. Among patients treated with adjuvant or neoadjuvant chemoradiation, the risk of distant recurrence is more significant than that of local recurrence⁴; therefore, the use of concurrent agents that have a proven effect on micrometastatic disease is more appealing.

HOW CAN IMPROVING CHEMORADATION RESPONSE IMPROVE OUTCOME?

Current Limitations of Chemoradiation in Locally Advanced Rectal Cancer

Based on recent data from EORTC and FFCD neoadjuvant trials as well as the Intergroup 0144 postoperative trial,^17,18,22 the addition of fluorouracil given as a protracted venous infusion or as a bolus with leucovorin leads to satisfactory (at least 90%) pelvic tumor control rates when radical surgery is used. Because most patients who receive 5-FU–based chemoradiation have very low risk of pelvic tumor recurrence, treatment intensification with biologic or cytotoxic agents is not likely to significantly improve pelvic control or overall survival, except possibly in the highest-risk patients. However, it is possible that improving the local treatment effect through novel means could eventually lead to a shift in the current paradigm and the increased use of organ preservation strategies in selected patients with locally advanced rectal cancer. If pathologic complete response rates reached 70% to 80%, as they have in anal canal carcinoma, the use of observation or full thickness local excision, with radical proctectomy reserved for salvage treatment, would be an attractive question to study in multi-institutional trials.

Organ Preservation in Rectal Cancer

Based on somewhat limited evidence, complete pathologic response of the primary tumor after chemoradiation (ypT0) for locally advanced rectal cancer correlates highly with microscopic disease sterilization in the pelvis. In a report from The University of Texas M. D. Anderson Cancer Center (MDACC), 9% (4/45) of patients with ypT0 disease had positive mesorectal lymph nodes; the positivity rate declined to 5% (2/41) when poorly differentiated tumors were excluded.³⁷ Another analysis of only clinically node-negative patients from MDACC (Table 1), combined with data from two other institutions, revealed a 1% (1/84) incidence of mesorectal nodal involvement among patients with cT3N0 tumors that had ypT0 disease after chemoradiation.³⁸ Patients undergoing full thickness local excision^39–42 or even observation⁴³ after neoadjuvant chemoradiation for T3N0 rectal cancer have had pelvic control rates as high or higher than expected with radical surgery. Four studies (Table 2) have reported rare tumor recurrences in patients undergoing full thickness local excision or observation instead of radical surgery, but the majority of patients in these series had excellent responses to chemoradiation.^39–42 In the first report of tumors treated with preoperative radiotherapy alone, 15/15 had T-stage downstaging.⁴⁰ In more recent reports of neoadjuvant chemoradiotherapy, pathologic complete response or microscopic residual disease (mRD) was seen in 17/17,³⁹ 11/11,⁴² and 23/26⁴¹ of tumors. Collectively, only two of the total 69 patients have been reported to have local tumor recurrence, both occurring in the MDACC series. These two patients had unfavorable characteristics; one had a clinically positive lymph node at presentation (radical surgery was not advised due to metachronous poor prognosis breast cancer), and the other had gross residual disease after chemoradiation.⁴¹

Table 1.

Correlation between pathologic response and probability of detecting positive mesorectal lymph-nodes among clinically staged T3N0 patients treated with preoperative (chemo)radiation followed by full thickness local excision.

Pathologic T Stage	Institution 1	Institution 2	Institution 3	Total
ypT0	0/27 (0%)	0/14 (0%)	1/43 (2%)	1/84 (1%)
ypT1	2/29 (7%)	0/12 (0%)	4/17 (24%)	6/58 (10%)
ypT2	15/95 (16%)	12/97 (12%)	4/60 (7%)	31/252 (12%)
ypT3	54/166 (33%)	62/164 (38%)	15/68 (22%)	131/398 (33%)
ypT4	0	5/5 (100%)	2/2 (100%)	7/7 (100%)

Data from University of Florida, University of Texas M. D. Anderson Cancer Center, and Washington University, St. Louis.³⁸

Table 2.

Studies evaluating neoadjuvant (chemo)radiation followed by transanal excision in T3 rectal cancer.

Study	No. Patients	% pCR/mRD	Pelvic Recurrence Rate (%) (5-yr actuarial)	Median Follow-Up
Mohiuddin, TJU, 199440	15	“Downstaged”	0	40
Kim, USF, 200139	17	100/0	0	19
Schell, UF, 200242	11	73/27	0	48
Bonnen, MDACC, 200441	26	54/35	6	51

Abbreviations: MDACC = The University of Texas M. D. Anderson Cancer Center; mRD = microscopic residual disease; pCR = pathologic complete response; TJU = Thomas Jefferson University; UF = University of Florida; USF = University of South Florida.

The most interesting example of an organ-preserving strategy in T3 rectal cancer is the prospective experience reported by Habr-Gama et al.⁴³ In 71 patients treated with neoadjuvant chemoradiation followed by observation who achieved complete clinical response, there have been no reported deaths due to rectal cancer with a median follow-up of 57.3 months. Two patients have undergone salvage abdominoperineal resection, but all are alive. Although these data are provocative, it is important to note that this experience has not been duplicated. However, it seems that radical surgery would not have improved outcome in these highly selected patients (collectively: 5 studies, 140 patients). Thus, it appears that exclusion of patients with poorly differentiated histology and, particularly, clinically enlarged lymph nodes leads to very high correlation between ypT0 status and microscopic disease sterilization in the mesorectum. These or similar criteria could be used in designing future trials evaluating this approach.

INCORPORATION OF MOLECULAR TARGETED THERAPY WITH RADIATION IN GASTROINTESTINAL TUMORS

Colorectal cancer is the first gastrointestinal cancer site for which targeted therapies have been shown beneficial. The monoclonal antibodies bevacizumab and cetuximab have changed the standard of care for patients with advanced disease and are being evaluated in the adjuvant setting. Both of these agents have proven radiosensitizing properties in the lab as well as the clinic. While the FDA has approved erlotinib and gefitinib, which are smallmolecule epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors, their roles in colorectal cancer have not been established. Thus, the study of monoclonal antibodies is currently more appealing.

Vascular Endothelial Growth Factor Inhibition

Randomized trials have demonstrated the efficacy of bevacizumab, a humanized antivascular endothelial growth factor (VEGF) monoclonal antibody. In the pivotal trial, bevacizumab was found to increase survival in patients with metastatic colorectal cancer when added to irinotecan, 5-FU, and leucovorin.⁴⁴ Bevacizumab also increased survival in patients with metastatic colorectal cancer when added to 5-FU and leucovorin.⁴⁵ Most investigations of targeted agents have focused on their benefits as components of systemic therapy for advanced disease, and studies in the adjuvant setting are under way in colon cancer. In addition, these molecular therapies may have important radiosensitizing effects. Mechanisms of radiosensitization are not yet defined but may include enhanced lethality of the endothelial cell⁴⁶ or tumor cell,⁴⁷ or improved vascular physiology leading to a reduction in tumor hypoxia.⁴⁸ The role of bevacizumab in preoperative rectal cancer therapy is under investigation. Willett et al evaluated bevacizumab and protracted venous infusional 5-FU with concurrent preoperative radiotherapy in a phase I trial.^49,50 Bevacizumab was well tolerated at the 5 mg/kg dose (n=6), whereas at 10 mg/kg, two patients had grade 3 gastrointestinal adverse events. The authors recommended the 5 mg/kg dose for further study. A complement of correlative studies showed that bevacizumab decreased tumor perfusion, interstitial pressure, and microvascular density in rectal cancers.⁴⁹ Final study results were available for 11 patients, and interestingly, 10 had clinical complete responses. Eight patients had microscopic residual disease and two had complete histologic responses. Compared with historic controls who received 5-FU–based chemoradiation, there appear to be fewer patients with gross residual disease and approximately the same number with complete histologic responses in the current study.⁴¹ This pattern of response is consistent with preliminary experiences that have not yet been formally presented (unpublished data, MDACC, personal communication, H. Safran, September, 2006), but inconsistent with at least one other experience.⁵¹ Several studies of preoperative radiotherapy with concurrent bevacizumab and 5-FU–based chemotherapy are ongoing.

The largest study of bevacizumab combined with radiation is a phase I dose escalation study in patients with locally advanced, unresectable pancreatic cancer conducted at MDACC. Forty seven patients received capecitabine (650–825 mg/m² twice daily) and bevacizumab in combination with radiation (50.4 Gy) to the gross tumor alone.⁵² Results demonstrated that bevacizumab is generally safe when combined with chemoradiation in this setting. Acute toxicity was minimal and easily managed with dose adjustments of capecitabine, without interruption or attenuation of bevacizumab or radiation doses. Bevacizumab did not appear to enhance acute toxicity; however, patients whose tumors invaded the duodenum appeared to be at higher risk for bleeding or perforation. Among the first 30 patients enrolled, there were three bleeding events associated with tumor invasion of the duodenum. After this was recognized, such patients were excluded from the study and no further bleeding events occurred among the final 16 patients enrolled. Overall, tumors in 9 (20%) of 46 evaluable patients had an objective partial response to initial therapy, including 6 of 12 tumors treated with a 5 mg/kg bevacizumab dose. Based on that trial, the recommended bevacizumab dose for further study is 5 mg/kg every 2 weeks with radiotherapy (50.4 Gy in 28 fractions) and concurrent capecitabine (825 mg/m² twice daily Monday through Friday).⁵²

Accrual has completed to an RTOG phase II trial of capecitabine-based chemoradiation with bevacizumab followed by systemic therapy with concurrent gemcitabine and bevacizumab (RTOG PA04-11). Ninety-four patients were treated and preliminary analysis indicates generally good tolerability in comparison to previous RTOG phase II studies with similar inclusion criteria. Exclusion of patients with tumor invasion of the duodenum seems to have avoided significant bleeding problems thus far. Initial safety analysis of 50 patients has revealed no tumor-associated duodenal bleeding. Further evaluations of bevacizumab as a radiosensitizer have been proposed in rectal cancer, lung cancer, and esophageal cancer.

Epidermal Growth Factor Inhibition

The EGFR signaling pathway represents another attractive target for biologic therapies. A randomized trial in patients with irinotecan-refractory colorectal cancer showed significant activity of the anti-EGFR antibody, cetuximab.⁵³ Cetuximab is also the only targeted agent that has been shown to improve local tumor control and overall survival outcome as a radiosensitizer, and is the only chemotherapeutic or targeted agent that has been approved by the FDA for use specifically as a radiosensitizer based on a phase III trial in patients with locally advanced head and neck cancer.⁵⁴ Patients were randomly assigned to receive definitive radiation alone or with concurrent cetuximab. Median durations of local tumor control were 14.9 months and 24.4 months in the two groups, respectively (hazard ratio for locoregional progression or death, 0.68; P = .005). With a median follow-up of 54.0 months, median overall survival duration was 49.0 months with combined therapy and 29.3 months with radiotherapy alone (hazard ratio for death, 0.74; P = .03).⁵⁴ Although similar efficacy improvements have been achieved with use of concurrent cisplatin in head and neck cancer, the toxicity increases in the cetuximab study were limited to rash and infusion reactions. Two single-arm studies have incorporated cetuximab with neoadjuvant chemoradiation regimens in locally advanced rectal cancer patients, both presented in abstract form.^55,56 Patients received standard doses of pelvic radiation (50.4 Gy) in combination with capecitabine (650–825 mg/m²)⁵⁵ or protracted infusional 5-FU (225 mg/m²/day),⁵⁶ plus cetuximab (400 mg/m² day 1, then 250 mg/m² weekly).^55,56 Pathologic complete response rates were approximately 12% in both trials. Given that neoadjuvant 5-FU–based chemoradiation produces pathologic complete response rates of 10% to 20%, it appears that the pathologic complete response rates from these trials do not provide a strong preliminary efficacy signal.

Receptor Tyrosine Kinase Inhibition

Use of receptor tyrosine kinase inhibition in combination with radiotherapy has been studied in patients with localized gastrointestinal malignancies. In a phase I dose escalation study from Brown University, patients with locally advanced pancreatic cancer received gemcitabine 75 mg/m² and paclitaxel 40 mg/m² weekly, and daily erlotinib, with 50.4 Gy radiotherapy to the primary tumor and regional lymphatics.⁵⁷ The maximum tolerated erlotinib dose was 50 mg/m². Median survival was 14.0 months and 6 (46%) of 13 patients had a partial response, suggesting that erlotinib-based chemoradiation regimens are worthy of further study in this disease.⁵⁷ A significant amount of high-grade diarrhea occurred, limiting the ability to administer full erlotinib doses. Toxicity could have been due to the concurrent gemcitabine and paclitaxel, the use of regional nodal irradiation, or the combination of radiotherapy and erlotinib.

Another phase I study ongoing at Memorial Sloan-Kettering Cancer Center is evaluating gemcitabine-based chemotherapy in combination with erlotinib. A phase I study at Duke University evaluated concurrent gefitinib (250 mg daily), capecitabine (650–825 mg/m² BID, 7 days/week) and radiotherapy in locally advanced pancreatic and rectal cancers.⁵⁸ Patients were treated to a dose of 50.4 Gy (45 Gy in 25 fractions to the regional lymphatics followed by an additional 5.4 Gy in three fractions to the primary tumor). Dose-limiting toxicity was seen in 2 of 6 patients with rectal cancer and 6 of 10 patients with pancreatic cancer. Diarrhea as well as nausea and vomiting with dehydration were common.⁵⁸ A recommended dose was not established due to these toxicities. These data do not suggest that gefitinib should be further studied with radiotherapy in gastrointestinal malignancies. None of the currently available EGFR inhibitors (erlotinib, gefitinib, and cetuximab) have been evaluated in completed multi-institutional trials in combination with radiation in locally advanced pancreatic cancer.

CONCLUSIONS

Neoadjuvant chemoradiation for patients with locally advanced rectal cancer has been proven superior to postoperative chemoradiation based on improvements in local control and sphincter preservation with lower toxicity rates. Patients with highrisk disease (T3 node-positive and T4 tumors) may benefit from the addition of novel radiosensitizers that enhance local treatment effect. On the other hand, selected patients with intermediate-risk disease (T3N0)—tumors that respond well to chemoradiation—appear to be candidates for organ-preserving strategies. The development of more active chemoradiation regimens would likely increase the pool of candidates for such an approach. Novel cytotoxic agents such as oxaliplatin and capecitabine could lead to improved local treatment responses and are currently being evaluated in the NSABP R-04 trial. However, greater improvement will likely be needed to achieve these objectives; incorporation of targeted therapies with radiation could lead to improved responses, local control, and possibly more widespread study of organ-preservation options in selected patients. The preliminary efficacy signals for both bevacizumab and cetuximab do not indicate an increase in disease sterilization rate. On the other hand, bevacizumab treatment may be associated with increased response of the gross tumor to microscopic residual disease. Further study of the receptor tyrosine kinase inhibitors gefitinib and erlotinib with radiotherapy should be done with caution due to possibly higher risk of severe diarrhea in patients with gastrointestinal disease.

What Questions Should Future Trials Ask?

Experience from the recent past informs us that phase III clinical trials will be increasingly difficult to conduct as the pace of new drug development increases in gastrointestinal malignancies and the landscape becomes more crowded. The design of clinical trials in general must take the future into account as much as possible. In the case of rectal cancer, the current de facto paradigm is that agents active in the first-line metastatic or adjuvant setting are the most appealing candidates for use with radiation. This could change in the targeted therapy era, but most likely, ongoing adjuvant colorectal cancer trials incorporating bevacizumab and cetuximab will have a profound effect on the design of the next phase III study with chemoradiation. Data from these trials will likely be available to coincide with the completed accrual of ECOG/Intergroup 3201 and NSABP R-04 trials. Future cooperative group studies incorporating novel agents with radiation in rectal cancer should focus on the highest risk patients with pelvic control as the primary end point. Meanwhile, it is reasonable to continue to explore the concept of organ-preserving strategies in selected locally advanced rectal cancer patients. This strategy could be optimized by incorporation of novel targeted agents and is currently feasible in single-institutional colorectal cancer study groups.