Gastrointestinal Cancers—Changing the Standard for Rectal Cancer and Establishing a New Standard for Liver Tumors

In this edition of Gastrointestinal (GI) Oncology Scan (1–5), we bid farewell to Dr Thomas Brunner and Dr Jennifer Wo, who have rotated off the editorial board. Dr Wo joined as one of the original associate editors in 2011, when Dr Zietman became editor-in-chief, and was instrumental in helping to define the standard set for the GI section of this journal. Dr Brunner joined the editorial board in 2013, and we quickly grew to appreciate the European and German perspective he brought to our group. We very much enjoyed working with both of them and are grateful for their services. At the same time, we welcome the addition of Dr Jim Murphy from the University of California, San Diego, and Dr Smith “Jim” Apisarnthanarax, from the University of Washington, both of whom have been providing outstanding reviews for this journal for many years. We look forward to adding their expertise and knowledge to our section.

For this edition, selection of articles was particularly challenging for our group, as there were quite a number of noteworthy publications in the recent literature. Rather than choose the difficult path, we opted to take the easy way out and included all of them. Highlighted here are 3 rectal cancer papers, 2 papers on radiation for hepatobiliary cancers, and 1 paper discussing long-term cognitive function in colorectal cancer patients. We hope you enjoy this edition.

Garcia-Aguilar J, et al. Effect of adding mFOLFOX6 after neoadjuvant chemoradiation in locally advanced rectal cancer: A multicentre, phase 2 trial. Lancet Oncol 2015. (6)

Summary:

In this multi-institution prospective phase 2 trial, patients with stage II or III rectal cancer within 12 cm from the anal verge were treated with preoperative chemoradiation. Radiation consisted of 45 Gy at 1.8 Gy/fraction followed by a boost to 5.4 to 9 Gy depending on the small bowel dose. Chemotherapy was delivered as 5-fluorouracil (5-FU), continuous infusion 225 mg/m², throughout radiation. Patients were assigned to 1 of 4 sequential treatment groups. In group 1, following completion of chemoradiation, patients underwent total mesorectal excision (TME) 6 to 8 weeks later. In groups 2, 3, and 4, patients received 2, 4, or 6 cycles, respectively, of modified oxaliplatin, leucovorin, and 5-fluorouracil (FOLFOX6) chemotherapy, followed by TME 3 to 5 weeks after the last cycle of chemotherapy. Accrual began in group 1 under the assumption of a 15% pathologic complete response (pCR) rate, with the added assumption that the pCR would increase by 5% in each successive group (Table 1).

Table 1.

Results from Garcia-Aguilar et al (6)

Variable	Group 1	Group 2	Group 3	Group 4	P value
Number of chemotherapy cycles	0	1.7	3.5	5	<.0001
Mean interval from CRT to surgery (wks)	8.5	11.1	15.4	19.3	.0001
Completion of all planned chemo cycles	NA	82%	81%	71%	.50
Grade 3–4 toxicity during FOLFOX	NA	4%	18%	36%	NS
pCR rate	18%	25%	30%	38%	.0036

Abbreviations: CRT = chemoradiation; FOLFOX = modified FOLFOX6 chemotherapy (oxaliplatin, leucovorin, and 5-fluorouracil); NA = not applicable; NS = not stated; pCR = pathologic complete response; wks = weeks.

In total, 292 patients were enrolled in all groups (71 in group 1; 74 in group 2; 71 in group 3; and 76 in group 4), and 259 patients were evaluable. There were no significant differences within the groups in terms of age, sex, performance status, T stage, N stage, overall stage, distance from anal verge, or tumor size. Staging by ultrasonography evaluation was more common in the earlier groups, whereas magnetic resonance imaging (MRI) was more commonly used in the later treatment groups (P=.004). Results for each group are shown in Table 1. On univariate analysis, group 4 had a significantly higher rate of pCR than group 1 (P=.028). Multivariate analysis, which included radiation dose, tumor size, stage, and distance from anal verge, showed that only the study group was associated with pCR rate (P=.048). Although the pelvic fibrosis scores given by the surgeon were higher in groups 2 to 4 compared with group 1 (P=.0001), the technical difficulty scores rated by the surgeon did not differ among the groups (P=.8). Finally, no differences in grade 3 or higher perioperative morbidity was seen among treatment groups.

Comment:

The German Rectal Trial established preoperative chemoradiation and TME 4 to 6 weeks later as the standard of care for locally advanced rectal cancer (7), but the role of adjuvant chemotherapy, although used in that trial, remains controversial (8–10). Despite the lack of strong data supporting its use, combination chemotherapy consisting of oxaliplatin with 5-FU or capecitabine is routinely given following preoperative chemoradiation and TME in practice. This trial by Garcia-Aguilar et al (6) was chosen because it highlights a simple yet potentially game-changing tweak to the accepted standard of care, which is to give chemotherapy during the rest period rather than wait until after surgery. The rationale for this alteration is threefold: (1) it addresses systemic disease sooner than the 2 to 4 months that would be needed to account for recovery from chemoradiation and TME; (2) it allows more time for maximizing pathologic response of the tumor to chemoradiation, given that the point of maximal response is likely beyond the standard 4 to 6 weeks used in the German Rectal Trial (11, 12); and (3) it consolidates the effects of chemoradiation to improve pCR rates. Concerns of delaying surgery have been related to the increased risk of operative morbidity due to radiation fibrosis and the potential for disease progression. The results presented in this study, which demonstrate the safety and potentially improved response with this approach, could have important ripple effects in the landscape for rectal cancer treatment. There has been interest in altering the surgical approach following a favorable response to chemoradiation, either with local excision (13, 14) or watchful waiting (15, 16). For such a strategy to be successful, optimizing the pathologic response is crucial, and taking advantage of chemotherapy is an important step toward achieving that goal. Watchful waiting protocols have already been using adjuvant chemotherapy after chemoradiation (15). This trial confirms the potential benefit that chemotherapy may add to the pathologic response.

Garcia-Aguilar J, et al. Organ preservation for clinical T2N0 distal rectal cancer using neoadjuvant chemoradiotherapy and local excision (ACOSOG Z6041): Results of an open-label, single-arm, multi-institutional, phase 2 trial. Lancet Oncol 2015. (17)

Summary:

American College of Surgeons Oncology Group (ACOSOG) study Z6041 represents a multi-institutional phase 2 single-arm trial that evaluated neoadjuvant chemoradiation therapy followed by local excision in 79 patients with T2N0 rectal cancer staged with either endorectal ultrasonography or endorectal coil MRI. Patients had tumors within 8 cm from the anal verge, measuring less than 4 cm, and involving less than 40% of the circumference of the rectum. Neoadjuvant chemoradiation therapy included capecitabine (825 mg/m², twice daily in weeks 1, 2, 4, and 5) and oxaliplatin (50 mg/m², in weeks 1, 2, 4, and 5). Radiation consisted of 45 Gy at 1.8 Gy/fraction followed by a boost of 9 Gy (54 Gy total dose). Surgery was either transanal excision or transanal endoscopic microsurgery and was performed 4 to 8 weeks after chemoradiation therapy. The authors found unanticipated high rates of toxicity during chemoradiation, and after 53 patients, they reduced the dose of capecitabine to 725 mg/m² and reduced the radiation boost dose to 5.4 Gy (total dose of 50.4 Gy).

Of the 79 enrolled patients, 29% had grade 3 GI toxicity during chemoradiation, including 34% within the original dose group to 19% within the revised dose group. At the time of surgery, 49% of patients had pCR (18). Four patients (5%) had pT3 tumors or positive margins after local excision and were recommended abdominoperineal resection (APR) according to protocol. At 3 years’ follow-up, 3 patients (4%) had developed local recurrence, and 2 of these patients underwent salvage APR. Disease-free survival at 3 years was 88%. One-year posttreatment quality of life did not decrease compared to baseline.

Comment:

APR with TME represents the standard of care for patients with clinically staged T2N0 rectal cancer within close proximity to the anal canal (19). This treatment leads to high cure rates (20) but also carries the consequence of a permanent colostomy as well as sexual, urinary, and other gastrointestinal morbidity. As a result, there has been great interest in studying sphincter-preserving procedures such as local excision in T2N0 rectal cancer. Prior research has found higher rates of local progression and possibly decreased overall survival (OS) compared to TME (21, 22). This phase 2 ACOSOG trial builds on local excision with the addition of neoadjuvant chemoradiation, and overall, the findings represent an important step toward defining a possible role for sphincter preservation among patients with T2N0 rectal cancer. The important findings include high rates of tumor response, relatively low rates of local recurrence, and stable quality of life. Although these results show promise, additional long-term follow-up will be needed to confirm these findings. The ACOSOG trial had moderate levels of toxicity during chemoradiation, but the regimen tested was not ideal. Since the inception of the ACOSOG study, other randomized trials have found that adding oxaliplatin to standard 5-FUebased chemoradiation did not improve outcomes but did increase toxicity (23–25). However, as in the trial by Garcia-Aguilar et al (6), the strategy of chemotherapy after chemoradiation could be integrated into the treatment algorithm to further improve outcome. Although the standard of care for low-lying T2N0 rectal cancer remains TME, the momentum for sphincter preservation is clearly gaining steam, and this study (17) provides further validity for this alternate treatment strategy, which for now may be reserved for patients who refuse or cannot tolerate APR. Further studies like this one will determine whether this approach can shift the present standard of care.

Appelt AL, et al. High-dose chemoradiotherapy and watchful waiting for distal rectal cancer: A prospective observational study. Lancet Oncol 2015. (26)

Summary:

In this prospective, multicenter study from Denmark, patients with resectable T2–3N0–1 adenocarcinoma of the low rectum were treated with concurrent chemotherapy and radiation therapy (RT). External beam RT consisted of 60 Gy to the primary tumor and 50 Gy to the elective lymph nodes, delivered with intensity modulated RT (IMRT) in a concomitant boost technique. A 5-Gy high-dose-rate brachytherapy boost was given to the tumor, using an endorectal applicator (guided either by clips or by MRI), in the last week of therapy. Chemotherapy was tegafur-uracil, 300 mg/m² daily, given orally. All treatment was completed in 6 weeks. Patients were regularly assessed every 2 weeks during treatment and at 6 weeks post-treatment, using a combination of endoscopy and imaging. Complete response was defined as no lesion on endoscopy, negative biopsy results, no nodal metastases on pelvic MRI, and no distant metastases on CT. Patients with incomplete response were recommended surgical resection, whereas those with complete response were observed using endoscopy every 2 to 3 months and using positron emission tomography–computed tomography (PET-CT) every 4 to 6 months over the first 2 years. The primary endpoint of the study was local recurrence (LR) at 1 year in the observation group. The accrual goal was 100 patients, using a modified Simon’s 2-stage approach with the premise that 1-year LR would be <30%.

Although the trial was closed prematurely due to slow accrual, 51 eligible patients underwent treatment, with 98% receiving RT according to protocol. In total, 40 patients (78%) had a complete response at 6 weeks posttreatment and underwent observation. Over a median follow-up period of 23.9 months, 9 of these 40 patients 22.5% had LR. The cumulative risk of LR after observation was 15.5% at 1 year and 25.9% at 2 years, with a median time to LR of 10.4 months. No patients had recurrence beyond 2 years of follow-up. The 2 year overall survival was 100% and the rate of local control with chemoradiation alone for the total trial population was 58%. Patient-reported functional outcome and physician-rated sphincter function were both good. At 1 year, 78% of the observed patients had experienced grade 1 rectal bleeding, but only 7% had grade 3 rectal bleeding.

Comment:

Similar to sphincter-preserving surgery for low-lying rectal cancer, as discussed above, nonoperative management is another alternative that continues to gain momentum, with more publications demonstrating very encouraging results. In this study by Appelt et al (26), patients with carefully selected, T2–3N0–1, low rectal cancers were treated with an aggressive multimodal regimen and were observed after complete clinical response 6 weeks after therapy. After concomitant boost IMRT and brachytherapy boost, 40 patients (78%) were able to achieve complete clinical response and were observed; 31 of these patients (78%) maintained local control over the duration of follow-up. Importantly, the overall survival at 2 years was 100%. This study presents another important step toward establishing watchful waiting as an acceptable alternative for low-lying rectal cancer following a complete clinical response. Although this study is relatively small in size, it is a well-designed prospective study that substantiates results from Dr Habr-Gama et al (15, 27) and others (16), suggesting that radical surgery can be safely avoided in a select group of patients after chemoradiation. Although further study is necessary to define the optimal RT regimen, improve our ability to identify patients with complete response, and determine the best surveillance regimen after therapy, the ball has clearly been set into motion with a prospective trial already in the United States evaluating neoadjuvant therapy followed by watchful waiting for clinical complete responderes (28), one that we anticipate will be a harbinger of many more like it to come.

Wahl DR, et al. Outcomes after stereotactic body radiotherapy or radiofrequency ablation for hepatocellular carcinoma. J Clin Oncol. (29)

Summary:

This single-institution retrospective study from the University of Michigan included 224 patients who were treated for hepatocellular carcinoma (HCC), using either radiofrequency ablation (RFA) or stereotactic body RT (SBRT) to compare freedom from local progression (FFLP), toxicity, and survival. The RFA group included 161 patients treated for 249 tumors, and the SBRT group had 63 patients treated for 83 tumors. The original treatment decisions were made at the discretion of the multidisciplinary tumor board. RFA was performed under ultrasonographic guidance with a 5-mm ablation cavity margin surrounding the tumor. Post-RFA imaging was used to confirm ablation effect 4 to 6 weeks after the procedure, with residual disease reablated. SBRT was given in either 3 or 5 fractions at the median dose of 30 Gy or 50 Gy (range: 27–60 Gy) prescribed to 99.5% of planning target volume (PTV). The median biological equivalent dose (BED) was 100 Gy. More than half of the patients were hepatitis C virus carriers.

Despite no significant differences in tumor size between the RFA group and the SBRT group (median: 1.8 cm vs 2.2 cm, respectively; P=.14), there were significant differences in the number of prior liver-directed therapies (0 vs 2, respectively), Child-Pugh scores (6.9 vs 6.2, respectively), and follow-up of living patients (50.9 vs 27.0 months, respectively). FFLP was defined as absence of progressive disease within or at the PTV margin by SBRT and the absence of recurrence within or adjacent to the ablation zone by RFA.

On univariate analysis, SBRT (97.4% and 83.8%, respectively) was superior to RFA (83.6% and 80.2%, respectively) in 1- and 2-year FFLP (hazard ratio [HR]: 2.63; P=.016). With tumor size stratification, there were no differences in tumor sizes <2 cm between the RFA and SBRT groups (P=.15). However, RFA was associated with significantly worse FFLP (P=.025) for tumor size ≥2 cm. On multivariate analysis, SBRT had improved FFLP to RFA (HR: 3.84; P=.002), while tumor size (P=.055) and prior treatments (P=.055) were not significantly associated with local progression. There were no differences in OS between RFA and SBRT patients. No differences in acute and late-grade 3+ adverse events were noted between the groups.

Comment:

There are no randomized trials that have tested external beam RT with other liver-directed therapies. As a result, RT has not been integrated into international treatment guidelines for HCC. One recent meta-analysis of 25 trials comparing transcatheter arterial chemoembolization (TACE) plus RT with TACE alone showed that TACE plus RT was associated with significantly better complete response and survival (30). This study by Wahl et al (29), which focused on HCC patients at relatively earlier stages (>90% T1/T2), showed that SBRT had significantly better FFLP than RFA, and this improvement was primarily seen in patients with tumor size ≥2 cm. RFA is recommended for small (<3 cm) tumors in most guidelines, in part because of the unsatisfactory control for larger tumors. In contrast, size-dependent tumor control has not been reported in SBRT series (31). Of note, differences in patient characteristics between RFA and SBRT in this study reflect the current clinical practice, including more prior liver-directed therapies in the SBRT group and more severe cirrhosis in the RFA group. SBRT is frequently chosen in patients with disease more refractory to other treatments, whereas RFA may be used to limit liver damage from its localized thermal effect. SBRT is often chosen also for more centrally located tumors and tumors adjacent to vascular structures or luminal organs. An important limitation of this study is that the SBRT group had shorter follow-up, so the possibility of late local failures must be considered. Also, other patterns of failure were not reported. Although combined RFA and TACE may be better than RFA alone (32), the results of this study are important in establishing a more important role of SBRT, particularly in larger hepatic tumors.

Tao R, et al. Ablative radiotherapy doses lead to a substantial prolongation of survival in patients with inoperable intrahepatic cholangiocarcinoma: A retrospective dose response analysis. J Clin Oncol. (33)

Summary:

This single-institution retrospective review reports outcomes and dose response analyses of 79 patients with unresectable intrahepatic cholangiocarcinomas (IHCC) treated at MD Anderson Cancer Center with definitive intent RT from 2002 to 2014. The study included patients with relatively large primary tumors (median: 7.9 cm), satellite intrahepatic lesions (39%), and nodal metastasis (58%). Nearly all patients (89%) received systemic therapy prior to RT. The majority of patients were treated with photon-based IMRT, whereas 32% were treated with passive scatter proton beam therapy. Median RT dose was 58.05 Gy (BED: 80.5 Gy) with various fractionation schemes (35–100 Gy in 3–30 fractions). From 2010 to 2014, patients were treated primarily with a dose-escalated approach, using a simultaneous integrated boost technique (n=19, 24% of patients).

In all patients, median survival was 30 months, and 3-year OS was 44%. Higher RT dose and BED were associated with superior OS and local control (LC). For patients receiving BED >80.5 Gy, 3-year OS was 73% and 3-year LC was 78% compared to 38% and 45%, respectively, for those receiving BED ≤80.5 Gy. On multivariate analysis, RT dose was the only factor associated with LC and OS when accounting for primary tumor size, satellite intrahepatic metastasis, and performance status. Nearly all local recurrences (89%) occurred within the high-dose region, and 67% of deaths were attributed to complications related to primary or metastatic intrahepatic disease progression. No patient developed radiation-induced liver disease, 4% had subacute cholangitis, and 9% had biliary stenosis.

Comment:

Although this study is not the first to show an RT dose-response relationship in hepatic malignancies (34–36), it is the largest study to date on the use of definitive RT in unresectable IHCC patients. These outcomes are comparable to resectable IHCC patients after curative intent surgery, which are encouraging given the more advanced disease in these patients. Correlation between RT dose and clinical outcomes as well as the apparent link between cause of death and intrahepatic disease progression lend additional support to the importance of local control and RT dose escalation in these patients. It is unlikely that dose escalation in this relatively rare disease will ever be tested in a randomized controlled setting. Although a BED of 80.5 Gy is likely not a “magic” threshold for LC and OS, these data do suggest that the more dose delivered to the primary tumor can result in better outcomes.

A few issues remain unresolved and should be mentioned. First, the true value of localized RT in the context of systemic therapy is still unclear in this disease. It is important to note that 89% of patients received chemotherapy prior to RT, which indicates that these were highly selected patients. Although the regimen of 67.5 Gy in 15 fractions (BED 97.9 Gy) that is being used as the target dose in the ongoing NRG-GI001 randomized trial of systemic therapy with or without localized RT appears justified based on this current study, it will be interesting to see how often this dose is able to be delivered and what the effect of dose de-escalation (down to 37.5 Gy) to meet organs at risk dose constraints will have. Second, selection of which patients should be receiving dose escalation needs further refinement. For example, are central tumors amenable to safe dose escalation? The weighting toward peripheral tumors in the dose escalation group and biliary stenosis of nearly 10% is noteworthy in this study. Furthermore, recent data suggest that dose constraints may need to be placed on the central biliary tree (akin to proximal lung airways) (37). Third, what is the role of concurrent chemotherapy in the setting of mildly hypofractionated RT? There may be room for improvement for LC even with dose escalation, particularly in light of the pattern of failures within the high-dose region described in this study. Fourth and finally, the optimal dose fractionation is still unknown. While this study used a variety of fractionation schedules, shortening the overall treatment time with hypofractionation could be considered both in improving LC and increasing patient convenience. Further studies testing hypofractioned regimens, including the NRG trial schedule as well as with SBRT approaches, will help to answer this question.

Vardy JL, et al. Cognitive function in patients with colorectal cancer who do and do not receive chemotherapy: A prospective, longitudinal, controlled study. J Clin Oncol. (38)

Summary:

This prospective trial sought to evaluate changes in cognitive function and underlying mechanisms in patients with localized colorectal cancer (CRC) (n=289; 60% of whom received chemotherapy), limited metastatic/recurrent CRC (n=73), and healthy controls (n=72). Participants completed cognitive assessments and questionnaires reporting cognitive symptoms, fatigue, quality of life, and anxiety/depression at baseline (before chemotherapy, if given) and at 6, 12, and 24 months. Correlative blood tests were drawn, including cytokines, clotting factors, apolipoprotein E genotype (shown to be a genetic marker for increased vulnerability to chemotherapy-induced cognitive decline) (39), and sex hormones. The primary endpoint was overall cognitive function.

Patients with localized CRC had more frequent cognitive impairment than healthy controls at baseline and 12 months (43%−46% vs 13%−15%, respectively) without significant effect of chemotherapy. The most affected domains of cognition were attention/working memory, verbal learning/memory, and complex processing speed. Cytokine levels were higher in CRC patients than in healthy controls. There were no associations among overall cognitive function and fatigue, quality of life, anxiety/depression, or blood tests. Also, age and education level did not predict cognitive dysfunction. Women had greater cognitive impairment than men, but men had a greater risk of cognitive decline over time. Patients who received chemotherapy were more likely to report cognitive symptoms; however, symptoms only weakly correlated with cognitive function. In summary, patients with CRC had significantly more impaired cognition than healthy controls, but there was no added effect of chemotherapy, implying that the mechanism of neurocognitive dysfunction is unknown. Rates of cognitive impairment ranged from 36% to 52% from baseline to 24 months compared to 6% to 19%, respectively, in healthy controls.

Comment:

Vardy et al (38) demonstrate the long-term effects of therapy and, perhaps, effects of the cancer itself in CRC patients. This study is notable for the rigorous and in-depth methods of cognitive testing (accounting for practice/learning effects of the tests themselves) and the relatively large number of patients that were included. Although prior studies had evaluated CRC patients (40, 41), they were neither as large nor as detailed in the scope of testing; the most detailed studies of neurocognitive impact of therapy have been conducted in breast cancer survivors (42, 43). Likewise, in a study evaluating long-term health outcomes in a British cohort in 2011, Khan et al (44) recognized that CRC survivors were at higher risk of dementia as a long-term health outcome than controls (HR: 1.68; 95% confidence interval: 1.20–2.35) (44).

This study, although a bit different than the usual studies we have highlighted in previous Oncology Scans (1–5), is an important reminder that the burden of long-term toxicity is not purely physical (eg, colostomy, chemotherapy-related neuropathy, and others) but may be neurocognitive as well. These investigators were not able to directly link chemotherapy as the main causative effect of cognitive impairment but did find that cytokine elevation, perhaps as a result of the tumor or therapy, were associated with this impairment. As with any longitudinal studies that rely on patient participation, patient dropout may have led to an underestimation of the rate of chemotherapy-related neurocognitive toxicity, as patients who did not complete the study had lower baseline cognitive scores, less education, and higher rates of English as a second language. It is hypothesized that those who struggled initially may have been less likely to complete assessments. This explanation may be the reason why there was a nonsignificant trend for more cognitive decline in patients with localized CRC who received chemotherapy than those who did not (32% vs 23%, respectively; P=.14).

This information presented by Vardy et al (38) about cognitive decline is relevant to all practitioners who care for cancer patients. When our long-term–follow-up patients complain of or demonstrate cognitive changes, we should not dismiss them as age-related decline, mental/emotional fatigue, or simply “chemo brain.” Appropriate testing in the acute setting and additional evaluation by other skilled providers should be considered. In attempting to compassionately and holistically care for our patients, we must be conscious of the prevalence of cognitive impairment. This study is an important reminder that cognitive impairment is a real consequence of having CRC, and it seems very reasonable to extrapolate these results to other cancers. In addition, further study on the possible contributing role of chemotherapy is currently underway (45).