Randomized Controlled Trial of Resection Versus Radiotherapy After Induction Chemotherapy in Stage IIIA-N2 Non–Small-Cell Lung CancerBackground: Induction chemotherapy before surgical resection increases survival compared with surgical resection alone in patients with stage IIIA-N2 non–small-cell lung cancer (NSCLC). We hypothesized that, following a response to induction chemotherapy, surgical resection would be superior to thoracic radiotherapy as locoregional therapy. Methods: Selected patients with histologic or cytologic proven stage IIIA-N2 NSCLC were given three cycles of platinum-based induction chemotherapy. Responding patients were subsequently randomly assigned to surgical resection or radiotherapy. Survival curves were estimated using Kaplan–Meier analyses from time of randomization. Results: Induction chemotherapy resulted in a response rate of 61% among the 579 eligible patients. A total of 167 patients were allocated to resection and 165 to radiotherapy. Of the 154 (92%) patients who underwent surgery, 14% had an exploratory thoracotomy, 50% a radical resection, 42% a pathologic downstaging, and 5% a pathologic complete response; 4% died after surgery. Postoperative radiotherapy was administered to 62 (40%) of patients in the surgery arm. Among the 154 (93%) irradiated patients, overall compliance to the radiotherapy prescription was 55%, and grade 3/4 acute and late esophageal and pulmonary toxic effects occurred in 4% and 7%; one patient died of radiation pneumonitis. Median and 5-year overall survival for patients randomly assigned to resection versus radiotherapy were 16.4 versus 17.5 months and 15.7% versus 14%, respectively. Rates of progression-free survival were also similar in both groups. Conclusion: In selected patients with pathologically proven stage IIIA-N2 NSCLC and a response to induction chemotherapy, surgical resection did not improve overall or progression-free survival compared with radiotherapy. In view of its low morbidity and mortality, radiotherapy should be considered the preferred locoregional treatment for these patients. |
Scalpels, Beams, Drugs, and Dreams: Challenges of Stage IIIA-N2 Non–Small-Cell Lung CancerApproximately 30% of patients who are newly diagnosed with non–small-cell lung cancers (NSCLCs) have locally advanced disease, i.e., stages IIIA and IIIB in the current staging system. Roughly 10% will be classified as stage IIIA-N2 on the basis of metastasis to the ipsilateral mediastinal lymph nodes. Furthermore, in a small proportion of patients, metastatic disease will be detected from primary tumor and lymph node specimens obtained during an operative procedure. Patients with stage IIIA-N2 NSCLC often have a good prognosis after surgery, but today, many will also be treated with postoperative adjuvant chemotherapy. In contrast, patients who have bulky mediastinal nodal involvement that is easily detected on a routine chest radiograph have poor prognosis after surgery alone. Today, however, N2 disease is often initially suspected when mediastinal node enlargement (i.e., >1.0-cm short-axis diameter) is detected on a staging chest computed tomography (CT) scan, but increasingly sophisticated imaging technologies, such as positron emission tomography (PET), integrated CT/PET, and endoscopic ultrasound (± fine-needle biopsy), have enhanced the detection of N2 disease. For the past 15 years or so, patients with stage IIIA-N2 NSCLC have been treated primarily with chemoradiotherapy. Nonetheless, because local (and distant) failure rates remain high, some lung cancer experts believe that selected IIIA-N2 patients and, in particular, those with so-called low-burden disease may benefit from surgery. To address the role of surgery after induction chemotherapy, the European Organization for Research and Treatment of Cancer (EORTC) conducted a randomized study (08941) of surgery versus radiotherapy following induction chemotherapy in patients with preoperatively detected stage IIIA-N2 NSCLC. The results of this important study are reported in this issue of the Journal. A review of the details of the EORTC 08941 is important, given the heterogeneity of IIIA-N2 NSCLC. In this study, all participating patients had cytologically or histologically proven, "unresectable" stage IIIA-N2 NSCLC. The definition of unresectable disease, although ultimately left to the discretion of the local surgeon, was carefully predefined in the EORTC protocol and is arguably in-line with standard practice. Preoperative staging consisted of a chest CT scan and an ultrasound or CT scan of the upper abdomen. More than 80% of the participating patients had mediastinal node involvement documented by mediastinoscopy, the gold standard for preoperatively confirming N2 disease. The remaining patients had disease pathologically documented at thoracotomy or via needle biopsy. Induction chemotherapy consisted of three cycles of a modern cisplatin-based regimen; most patients received either gemcitabine or paclitaxel with cisplatin. Only patients who demonstrated at least a "minor" response were subsequently randomly assigned to radiotherapy or to surgical resection. Thus, the "best of the best" patients were selected as candidates for surgery after induction therapy. Mediastinal "downstaging" was defined as the absence of tumor at pathologic examination of the resected mediastinal nodes. Postoperative radiation therapy was recommended in patients with an incomplete resection or in those with unresectable disease. The primary study endpoint was overall survival. The study enrolled 582 patients; 87% received the planned three cycles of induction chemotherapy, and the overall response rate was 61%. Ultimately, 332 (57%) patients were randomly assigned to surgery or radiotherapy. Of the 167 patients randomly assigned to the surgical arm, 92% underwent an operative procedure; 50% of the procedures were complete (R0) resections and 47% of the procedures were pneumonectomies. Ninety-three percent of the 165 patients in the radiotherapy arm were irradiated; compliance with the planned radiotherapy prescription was acceptable. Median (16.4 versus 17.5 months) and 5-year survival rates (15.7% versus 14%) were very similar in the two arms. Surgery was associated with less local regional relapse but greater extrathoracic recurrence, whereas the opposite relapse pattern was observed with radiotherapy. Notably, these results are remarkably similar to those reported for an aborted Radiation Therapy Oncology Group trial that was closed due to slow accrual but also to allow the activation of the recently completed North American Intergroup study (INT-0139). The North American Intergroup study examined the utility of induction chemoradiation therapy (i.e., trimodality therapy) followed by surgery compared with chemoradiotherapy alone (i.e., bimodality therapy). Whether surgery following induction chemoradiation therapy (i.e., trimodality therapy) will improve overall survival remains to be determined, but initial reports suggest that it does not. The results of EORTC 08941 have obvious implications for the treatment for patients with stage IIIA-N2 NSCLC that was preoperatively detected and pathologically confirmed. Accordingly, these data will be thoroughly parsed by surgeons, radiation oncologists, and medical oncologists alike to assess how each specialty's contribution might be improved and the negative aspects minimized. Surgeons will justifiably fret over the high number of pneumonectomies performed in this trial because previous experience has shown a strong negative outcome with pneumonectomy after induction chemotherapy. Indeed, important surgical lessons have been learned from previous combined modality trials during the past 20 years that have led to better patient selection and lower morbidiy and mortality. Thorough preoperative evaluation of cardiopulmonary function, preoperative and intraoperative assessment of the extent of residual N2 disease, and the potential for complete (R0) resection and the use of parenchyma-sparing resections (lobectomy ± bronchovascular sleeve resection) are crucial to achieving better surgical outcomes and optimal local control. The results of EORTC 08941 further validate these earlier observations and emphasize the importance of performing pneumonectomies only in highly selected patients if at all. Surgeons might also advocate pursuing a strategy of performing resection only on patients who experience "mediastinal downstaging", because this subset of tumors may behave more like de novo stage I or II disease. This may be true, but as far as we know now, resection in this setting is simply diagnostic and prognostic and not of proven therapeutic benefit. As the authors note, it will take a properly designed prospective trial to address this possibility, but given the time it took to complete this and similar induction therapy studies, such a trial is not likely feasible. It is nonetheless worth noting that local control was better in the surgical than the radiation arm of the trial. Local recurrence of lung cancer within an irradiated field is a highly morbid event that can affect quality of life in the latter stages of disease. Thus, patients who are good candidates for surgery may still be appropriately managed by resection rather than radiation. Careful interdisciplinary planning of patient care at the time of diagnosis to optimize selection of the local control modality is essential to achieving the best outcome in this patient population. Radiation oncologists will claim that continued improvements in the delivery of radiation therapy may produce even better local control than that achieved in the radiation therapy arm of EORTC 08941. In fact, radiation oncologists now define targets more conservatively by treating only the nodes clearly involved by CT or PET scan, thereby avoiding irradiation to the adjacent structures: lung, esophagus, heart, and spinal cord. Modern techniques, such as intensity-modulated radiotherapy or other sophisticated technologies (tomotherapy, cyberknife, protons) might also provide more precise targeting and allow higher radiation dosage in conjunction with chemotherapy. The higher radiation doses permitted by these techniques could lead to improved local control and improved overall survival. Simultaneously, reducing target volumes would decrease toxicity. But such an assumption requires prospective evaluation. Measures that shorten overall treatment time, such as accelerated fractionation schemes, also merit further consideration. In theory, this might obliterate surviving clonogenic tumor cells that can proliferate when there are gaps or prolongation in treatment time. Unfortunately, the attendant-increased normal tissue toxicity observed in some studies does not bode well for combining this approach with concurrent chemotherapy. To date, these accelerated methods have been accomplished without technologic advantages and perhaps would be somewhat less toxic if used. Even when improved technology is used, however, multiple daily radiation treatments are unpopular, and in the end, delivering the radiation dose in a shorter time may be more important than using multiple daily fractions. The main challenge facing the medical oncologist is an all too familiar one; namely, how to reduce the high rate of systemic recurrence, a consequence of the disappointing activity of extant chemotherapy regimens. Less imaginative medical oncologists will likely fall back on the time-honored tradition of tweaking the chemotherapy recipe, but this may be fool's gold. One of the more creative aspects of the EORTC trial was the embedding of three phase 2 studies within the larger phase 3 effort. Although not a direct head-to-head comparison of the individual chemotherapy regimens, the data derived from this series of embedded trials strongly suggest that commonly used platinum-based doublets produced similar results. What modest differences exist are likely a result of unrecognized biologic differences in the treated populations and are probably of little clinical relevance. The message in this unique experience is that further efforts to empirically hone the induction chemotherapy regimen using existing cytotoxic agents are not likely to prove hugely beneficial. Parenthetically, the introduction of so-called targeted agents to induction regimens requires careful patient selection. The incorporation of these drugs should be data and hypothesis driven and not simply empiric. In short, the EORTC 08941 data indicate that chemoradiation therapy remains an appropriate therapeutic strategy for the subset of IIIA NSCLC patients with preoperatively detected N2 disease. The results also emphasize the importance of careful patient selection for surgery and of the type of lung resection. Ideally, such surgery should be performed by an experienced thoracic surgeon who is practicing in a facility that performs a high volume of such procedures. In the end, EORTC 08941 generates as many questions as it answers. Will parenchymal sparing surgery following induction chemotherapy prolong survival? Is surgery an option for patients with mediastinal downstaging? The optimal radiotherapy fractionation scheme, treatment volume, and tumor targeting all require better definition. Is there an "optimal" induction chemotherapy regimen, should "consolidation" chemotherapy be used, what is the optimal dosing of chemotherapy with concurrent radiotherapy, and how do we incorporate newer targeted therapies into induction protocols? Each of these issues is important, and isolating their effect within well-designed clinical trials will be challenging. As we move forward, it is our dream to also focus on prospectively validating putative molecular markers of prognosis, drug sensitivity, and resistance. Hopefully, these promising technologies can be used to guide patient selection and treatment decisions in the future. |