Chemoradiation: Is There a Role?

One of the primary questions in locally advanced pancreatic cancer is the role that radiation plays and the importance of locoregional control. A postmortem study from Johns Hopkins indicated that up to 30% of patients with pancreatic cancer died of “locally destructive” rather than extensive metastatic disease, highlighting the potential importance of achieving locoregional control in this malignancy.

There are conflicting data regarding the role of combined-modality therapy for locally advanced pancreatic cancer. In terms of concurrent administration of chemotherapy with radiation, an early study conducted by the Gastrointestinal Tumor Study Group (GITSG) demonstrated the superiority of chemotherapy (bolus 5-fluorouracil [5-FU]) administered together with radiation when compared with radiation alone in patients with locally advanced disease, with improved median survival, and with 1-year survival rates. Although this informs the current practice standard of administering radiosensitizing doses of chemotherapy concurrently with radiation, it is worth noting that a much later published study conducted by the Eastern Cooperative Group (ECOG) failed to demonstrate any survival improvement when mitomycin C (MMC) and 5-FU were administered concurrently with radiation compared with radiation alone.

The Question of Timing: Induction Chemoradiation

Most studies evaluating a treatment paradigm that includes chemoradiation for locally advanced disease have used it as part of initial therapy ( Table 1 ). Several of these have used a randomized design to analyze whether sequential chemoradiation, followed by chemotherapy, is superior to chemotherapy alone (ie, what benefit is conferred by the addition of induction chemoradiation in this disease context). Notably, these trials have produced markedly discordant results: 2 trials demonstrated a survival advantage, 1 trial showed no difference, and 1 trial actually indicated a detrimental effect of using induction chemoradiation in this patient population. These disparate findings may be potentially explained by several factors that could have affected clinical outcomes: different radiation doses and treatment plans, radiosensitizing agents, and stand-alone chemotherapy regimens. For example, in the one modern phase III study to demonstrate a shorter median survival with combined-modality therapy, conducted by the Federation Francophone de Cancerologie Digestive (FFCD) and the Societe Francophone de Radiotherapie Oncologique (SFRO), induction chemoradiation consisted of 6000 cGy of radiation over 30 fractions with 2 concurrent cycles of cisplatin (20 mg/m ² /d on days 1 through 5 during weeks 1 and 5) and continuous infusion 5-FU (300 mg/m ² /d, days 1 through 5 for 6 weeks). Not surprisingly, patients receiving this chemoradiation regimen incurred more frequent grade 3 to 4 toxicities when compared with those receiving chemotherapy alone, both during the corresponding induction periods and during the maintenance chemotherapy phase with gemcitabine, leading to early treatment interruption and shorter duration and lower cumulative dose of subsequent chemotherapy. Such factors may have contributed to the poorer outcomes, highlighting the importance of radiation treatment planning and selection of radiosensitizing agents, as is discussed later.

Delayed Chemoradiation Following Induction Chemotherapy

An emerging and promising treatment paradigm for locally advanced pancreatic cancer consists of delaying chemoradiation until after patients have completed a period of induction chemotherapy. This strategy would address 2 important issues: (1) the need to optimally address systemic disease first, which is usually the most significant factor affecting longevity in most patients; and (2) limiting the use of radiation to that subgroup of patients whose tumors are well-controlled with this initial period of systemic therapy and do not develop metastatic disease. This approach would spare a subset of individuals the potential toxicities associated with radiation.

Support for this strategy came from the Groupe Cooperateur Multidisciplinaire en Oncologie (GERCOR), which performed a retrospective analysis on 181 patients with locally advanced pancreatic cancer who had participated in several separate phase II or phase III trials and received initial chemotherapy with a gemcitabine-based combination regimen for 3 months or more. For those patients who showed good disease control (70.7%, n = 128), the treating investigator was then permitted to decide whether to continue with the same chemotherapy or proceed onto consolidative chemoradiation. Seventy-two (56%) patients received consolidative radiation (5500 cGy with concurrent infusional 5-FU), whereas the remaining 56 (44%) patients continued with chemotherapy alone. Patients receiving chemoradiation had longer rates of overall survival (OS) and progression-free survival (PFS) (median OS 15.0 vs 11.7 months, P = .0009; median PFS 10.8 vs 7.4 months, P = .005).

Krishnan and colleagues similarly performed a retrospective analysis of 323 patients with locally advanced pancreatic cancer treated at M.D. Anderson Cancer Center, 76 of whom received gemcitabine-based induction chemotherapy (for a median duration of 2.5 months) before chemoradiation. These patients, when compared with those who had received chemoradiation as part of their initial treatment, were found to have significantly longer OS and PFS rates (median OS, 11.9 vs 8.5 months, P <.001; median PFS, 6.4 vs 4.2 months, P <.001, respectively).

Although intriguing, the retrospective nature of both these analyses prevents any definitive conclusions about the benefit of delayed chemoradiation compared with either no chemoradiation or induction chemoradiation. Data are now available from several prospectively designed phase II studies, and most recently, a large randomized phase III study, that have analyzed this sequential approach of induction chemotherapy followed by chemoradiation ( Table 2 ). These studies have highlighted several important findings: (1) approximately 13% to 39% of patients receiving induction chemotherapy will not be candidates for subsequent chemoradiation, most commonly due to disease progression; and (2) patients who successfully undergo sequential chemotherapy followed by chemoradiation typically have survival outcomes greater than 1 year.

More recently, this strategy of delayed chemoradiation was attempted to be addressed more definitively via a phase III European study conducted by GERCOR (LAP-07). In this trial, patients with locally advanced pancreatic cancer were initially randomized to receive gemcitabine with or without the epidermal growth factor receptor inhibitor erlotinib for 4 months. Patients who did not progress during this initial treatment were then included in the second randomization step, which involved either continuation of systemic therapy for 2 more months (patients receiving erlotinib could be maintained on this agent indefinitely) versus chemoradiation (54 Gy in 30 fractions, with concurrent capecitabine at 800 mg/m ² twice a day). Results of this 442-patient study were initially presented by Hammel and colleagues at the 2013 annual American Society of Clinical Oncology meeting. Of 442 patients enrolled to this study, 269 reached the second randomization step, representing a 39% dropout rate following induction chemotherapy (primarily due to progressive disease). Updated efficacy results in 2014 indicated that patients assigned to receive chemoradiation experienced no improvement in survival compared with those continuing with chemotherapy alone (median OS, 15.2 vs 16.5 months, respectively; log-rank P = .829), although there was a trend toward improved PFS in the chemoradiation-treated patients (median PFS 9.9 vs 8.4 months, log-rank P = .055). Not surprisingly, patterns of disease progression differed between the 2 arms: patients receiving chemoradiation did have a lower rate of subsequent local progression as well as a longer treatment-free interval before having to resume therapy at a later time. As the systemic therapy used in LAP-07 (gemcitabine ± erlotinib) is no longer commonly given, it is unclear whether the results of this trial will still be applicable, and whether newer and more effective modern chemotherapy regimens administered in the induction setting will attenuate or magnify any potential benefits associated with radiation in this patient population.

Selection of Chemotherapy Administered Independently of Radiation

Two regimens have emerged as front-line standards of care for the treatment of metastatic pancreatic cancer: FOLFIRINOX (a biweekly regimen consisting of 5-FU administered over 48 hours plus leucovorin, irinotecan, and oxaliplatin); and the combination of gemcitabine and nab-paclitaxel. Both of these regimens demonstrated significantly prolonged survival times compared with gemcitabine monotherapy in randomized phase III trials (FOLFIRINOX vs gemcitabine, median OS 11.1 vs 6.8 months; hazard ratio [HR] 0.57, P <.001; gemcitabine/nab-paclitaxel vs gemcitabine, median OS 8.5 vs 6.7 months; HR 0.72, P <.001), as well as improvements in other clinically relevant endpoints including PFS and objective response rate. However, it is important to recognize that both of these trials enrolled patients exclusively with metastatic disease, and hence, use of either of these regimens in the locally advanced setting requires some extrapolation as the experience to date in this context has been limited to small single-institution series. Of note, NCCN guidelines do not make a distinction in choice of chemotherapy between patients with metastatic versus locally advanced disease. In a recently opened national cooperative group trial (RTOG 1201), all patients will receive induction chemotherapy with the combination of gemcitabine and nab-paclitaxel, followed by a randomization to 1 of 3 arms: continued chemotherapy, standard-fractionation radiation, or higher-dose intensity-modulated radiation ( ClinicalTrials.gov NCT01921751 ).

Selection of Radiosensitizing Agents

Historically, most studies for locally advanced pancreatic cancer have delivered fluoropyrimidine- or gemcitabine-based therapies concurrent with radiation. Older trials used bolus 5-FU, but administration of fluoropyrimidine therapy continuously throughout the course of radiation (either as infusional 5-FU or as the oral prodrug capecitabine), rather than in bolus fashion, may provide superior radiosensitization and represents the currently accepted standard of care. Although gemcitabine is a more potent radiosensitizer, rates of severe toxicity may be higher with gemcitabine-based chemoradiation, as suggested from a retrospective analysis by Crane and colleagues of patients with locally advanced disease treated at M.D. Anderson. However, other data have shown that full doses of gemcitabine, if delivered with limited volume and highly conformal radiotherapy, can be well-tolerated and efficacious. A comparison of fluoropyrimidine-based versus gemcitabine-based chemoradiation was performed in the context of a randomized phase II study (the SCALOP trial), in which 74 patients with locally advanced disease were randomized, after a 12-week course of induction chemotherapy with gemcitabine/capecitabine, to receive radiation (5040 cGy over 28 fractions) with either concurrent capecitabine (830 mg/m ² twice a day Monday through Friday) or gemcitabine (300 mg/m ² weekly). Median OS was significantly superior in the former group (15.2 vs 13.4 months; adjusted HR 0.39, P = .012), suggesting that a capecitabine-based regimen may be the preferred choice in the context of consolidative chemoradiation.

Other radiosensitizing agents that have been explored for locally advanced pancreatic cancer include new oral fluoropyrimidine derivatives (S-1 ), taxanes, epidermal growth factor receptor inhibitors (erlotinib, cetuximab ), anti-angiogenic therapy (bevacizumab ), and protease inhibitors (nelfinavir ); however, at present, these should only be administered in the context of a clinical trial. One interesting recent study intending to enhance the effects of radiation involved localized injection of TNFerade, a replication-deficient adenoviral vector expressing the transgene tumor necrosis factor-α, regulated by the radiation-inducible promoter Egr-1. However, final results of a 187-patient randomized phase III trial demonstrated that the addition of TNFerade to 5-FU-based chemoradiation did not prolong survival compared with standard of care therapy (median OS 10.0 vs 10.0 months; HR 0.90, P = .26).

Treatment Volumes

The earlier landmark trials of chemoradiation in locally advanced pancreatic cancer treated large volumes covering the pancreatic tumor and draining lymph nodes with a 2-field technique, with concurrent chemotherapy. Because of the toxicities associated with combined radiotherapy and chemotherapy, the treatment course consisted of 2 weeks of radiotherapy, followed by a 2-week recovery period, followed by 2 more weeks of radiotherapy. Despite this suboptimal method of delivering radiotherapy, these studies demonstrated a significant benefit in favor of radiotherapy.

More recent data suggest that reduction in these treatment volumes improves the tolerability of radiotherapy. Multiple series using limited volumes including only gross disease, without elective nodal coverage, show improved toxicity rates and no compromise in local progression rates. Avoiding prophylactic nodal irradiation may allow intensification of therapy without exceeding normal tissue constraints. Murphy and colleagues were able to deliver radiation with concurrent full-dose gemcitabine by treating gross tumor volume (GTV) plus 1-cm margin only and found that only 5% of these patients experienced peripancreatic lymph node failures. These data are compelling, indicating that prophylactic nodal irradiation may be omitted in certain clinical scenarios.

With more focal radiation fields, target volumes as well as organs at risk must be accurately delineated. High-quality biphasic (early arterial and portal venous phase) CT and fluorodeoxyglucose (FDG)-PET imaging are both useful in helping to identify pancreatic tumors and to distinguish tumor from adjacent organs and vessels. In addition to intravenous contrast, dilute oral contrast can be used to define the pancreatic head better against the duodenum.

Respiratory-Motion Management

More conformal treating planning requires consideration of target motion. Pancreatic tumors, as with any upper abdominal organ, are subject to respiratory motion. The magnitude of motion for pancreatic tumors has been shown to be more significant in the cranio-caudal direction and can be as much as 2 to 3 cm from inspiration to expiration. Four-dimensional (4D) CT should be obtained at simulation to characterize this tumor motion during the respiratory cycle and subsequently construct an internal tumor volume (ITV). If the tumor motion is significantly great, construction of a large ITV can result in much larger treatment fields and potentially increased toxicity. Given this, multiple strategies can be used to reduce the expansion of treatment volume without compromising tumor coverage. Respiratory gating may be used to limit treatment to only specific phases of the respiratory cycle whereby tumor motion is minimal or during a phase when the overlap between planning treatment volume (PTV) and organ at risk is minimal. Mechanical techniques, such as abdominal compression to limit diaphragmatic excursion and active breathing control breath-holding technique, can also be used to minimize respiratory and tumor motion.