From Radiotherapy to Targeted Therapy: 20 Years in the Management of Non-Small-Cell Lung Cancer
From Radiotherapy to Targeted Therapy: 20 Years in the Management of Non-Small-Cell Lung Cancer
ABSTRACT: Non-small-cell lung cancer (NSCLC) is the leading cause of cancer death worldwide. Before 1980, radiotherapy was considered the only real recourse in advanced disease. In 1995, a landmark meta-analysis of trials conducted in the 1980s and early 1990s demonstrated a survival benefit with platinum-based chemotherapy. Newer chemotherapy agents and improved supportive care measures have allowed more patients to benefit from chemotherapy with reduced toxicity. Concurrent platinum-based chemotherapy and radiotherapy has improved the survival in stage III disease, and recently chemotherapy has also demonstrated improved survival in resected early-stage disease. The majority of patients still present with advanced unresec disease for whom the prognosis remains poor, but for key subpopulations the outlook has improved markedly since the emergence of targeted therapies directed against the epidermal growth factor receptor and vascular endothelial growth factor receptor pathways. Patient selection and the incorporation of targeted therapies with cytotoxic chemotherapy are the focus of many ongoing studies, and there is an abundance of new agents undergoing clinical trials. Together, these developments have moved us away from the nihilism of 20 years ago into an era of unprecedented optimism in taking on the many remaining challenges of managing NSCLC in the 21st century.
Lung cancer is the leading cause of cancer death worldwide, and it has been the most common cause of death from malignancy in the United States in men since the mid-1950s and in women since the mid-1980s. While incidence and death rates from lung cancer peaked in men in 1990, the incidence and death rates in women continue to increase. Approximately 85% of lung cancer cases are non-small-cell lung cancer (NSCLC), with the dominant histologic subtype in women drifting from squamous carcinoma to adenocarcinoma during the mid-1980s and the same drift occurring in men during the mid-1990s. In the past 20 years, our methods of detecting, defining, and treating NSCLC have all undergone considerable change. This paper reviews the most significant of these changes and assesses their overall impact on the clinical management of NSCLC.
Diagnosis: Screening, Staging, and Histologic Definition
In those who are eligible, surgery has always offered the best chance of cure. A number of studies of screening in high-risk individuals by chest x-rays, sputum cytology, and computed tomography (CT) scanning have been reported since the mid-1980s, but the benefits and cost-effectiveness of routine screening for lung cancers have still not been definitively established. The majority of patients with NSCLC continue to present with advanced inoperable disease.
Apart from overt metastatic disease, the nodal status of NSCLC remains the most common determinant of operability. Up to approximately 70% of lymph nodes found to be positive at thoracotomy are in stations not accessible by standard mediastinoscopy. Although the newest video-assisted thoracoscopic surgical (VATS) procedures potentially allow access to all lymph node stations, the real debate in recent years has become when and in whom should a mediastinoscopic assessment be performed.
Because of the high false-negative rate and moderate false-positive rate for CT scanning alone in the mediastinum, some have recommended preoperative mediastinoscopy in all potentially operable patients regardless of the CT scan results, while others routinely omit this procedure for small peripheral lesions without enlarged lymph nodes. Positron-emission tomography (PET) or PET/CT scanning is becoming commonplace and appears to be complementary to CT alone, with the primary role of the combination being in its high negative-predictive value. A negative PET/CT work-up may eventually preclude preoperative mediastinoscopy, with invasive preresection staging being restricted to confirming/denying suspicious lymph nodes, because of the moderate false-positive rate associated with both imaging modalities.
The American Joint Committee for Cancer Staging (AJCC) originally adapted the tumor-node-metastasis (TNM) system of Denoix into a staging system for NSCLC in 1974. Since then, subgroups within the initial stages have been recognized (ie, with different outcomes or likely to benefit from different interventions), and two revisions of the AJCC system have been issued.[8,9]
The World Health Organization system for histologic classification of lung cancers has also undergone a series of revisions since its first publication in 1967. Between 1981 and 1999, categories covering large-cell neuroendocrine carcinoma, spindle/giant cell carcinoma, and adenocarcinoma with mixed histologies were all added.[10,11] In 1999, bronchioloalveolar carcinoma, listed as a subtype of adenocarcinoma since 1967, was more rigidly defined to include only tumors demonstrating pure lepidic, noninvasive growth.
Treatment of Potentially Operable Disease
Improved preoperative work-up for predicting residual lung function, improved surgical techniques/equipment, and better supportive care have all made some previously inoperable patients operable. With additional improvements in imaging to fully delineate the extent of disease, minimalist surgery has also become more practical. VATS lobectomies have been reported as equivalent to conventional lobectomies, while wedge resections or segmentectomies have been associated with higher rates of local recurrence than lobectomies.[12,13] Probably the most significant change in the operative arena, however, has been in clarifying the roles of chemotherapy and radiotherapy as adjuncts to surgery, such that the standard management of any potentially operable NSCLC should now include a multidisciplinary team discussion.
Adjuvant radiotherapy, as a means to sterilize the resection margins, tumor bed, and regional lymph nodes postoperatively was demonstrated to dramatically reduce the risks of local recurrence from 41% to 3% for node-positive squamous carcinoma of the lung in the mid-1980s. However, adjuvant radiotherapy was dealt a heavy blow in 1998 by the postoperative radiotherapy (PORT) meta-analysis. Not only did PORT not show any survival benefit in completely resected N2 disease, but survival in N0-1 disease was worsened, with the survival curves separating after approximately 4 months postradiotherapy. This timing fits well with radiation-induced pneumonitis as the potential cause. Many believe that these results cannot be applied to radiotherapy using more modern planning techniques or, at the very least, that the question should be addressed for such techniques within more up-to-date trials of PORT.
In 1995, a comprehensive meta-analysis suggested that alkylating agents given in the adjuvant setting adversely affected survival (hazard ratio [HR] = 1.15, P = .005). In contrast, cisplatin-based adjuvant trials showed a trend towards benefit (HR = 0.87, P = .08). As a consequence of this meta-analysis, a series of large randomized, controlled adjuvant trials focusing on platinum-based regimens were undertaken.
• Key Trials-The first such trial, the Adjuvant Lung Project Italy (ALPI), did not show a significant difference for adjuvant MIC chemotherapy (mitomycin, ifosfamide, cisplatin) in stage I-IIIA disease in terms of overall survival (HR = 0.96, 95%confidence interval [CI] = 0.81-1.13). Soon afterwards, the positive results of the International Adjuvant Lung Cancer Trial (IALT), using a variety of more modern cisplatin-based doublets in stage I-III disease, were reported. Overall survival favored adjuvant chemotherapy (HR = 0.86, 95% CI = 0.76-0.98), translating to a 4.1% increase in 5-year survival. Despite concerns about the impact of imbalances in PORT between the arms, and the lower-than-planned recruitment numbers, the result was hailed as a new standard of care. The study appeared to show more benefit for patients with higher-stage disease than for those with lower-stage disease, although the investigators were resistant to any formal subgroup analysis.
The following year, the positive results of the JBR.10 trial of cisplatin/vinorelbine in stage IB and II disease (HR = 0.69, 95% CI = 0.52-0.91) and the interim results at 34 months' follow-up of the CALGB 9633 trial of carboplatin/paclitaxel in stage IB disease (HR = 0.62, 95 %CI = 0.41-0.95) were both announced at the 2004 American Society of Clinical Oncology (ASCO) meeting.[19,20] Subgroup analysis within JBR.10 showed that only patients with stage II (HR = 0.59, 95 %CI = 0.42-0.85), and not stage IB disease, benefited from adjuvant chemotherapy.
At ASCO 2005, the results of yet another positive trial, in stage IB-IIIA disease were announced, with subset analysis again suggesting benefit from chemotherapy in stages II-IIIA, but not stage IB. Some may initially have speculated that the cisplatin doublets used in these other trials were not as effective as the carboplatin/paclitaxel used for stage IB in CALGB 9633. However, at ASCO 2006, when the updated CALGB 9633 results at 54 months' median follow-up were announced, the trend toward improved survival was no longer statistically significant (HR = 0.80, 90% CI = 0.60-1.07). Whether the negative result of CALGB 9633 reflects drug effects, stage effects, or simply the fact that, due to slow accrual, only 344 patients were randomized (making it the smallest of these key adjuvant studies and therefore the lowest powered) continues to be debated.
• Conclusions-We are left in a position where the evidence base does not support the widespread use of adjuvant chemotherapy for stage I disease. Stage IB patients certainly require further study in the context of a clinical trial. With regard to the optimal adjuvant chemotherapy regimen for resected higher-stage disease, the question remains open as to whether carboplatin/paclitaxel (the US standard in advanced disease) should be used in this setting in the absence of data, or whether the fact that more data are available for cisplatin in combination with vinorelbine or other agents should alter our practice.
It is now likely that we will have to follow a similar path to our colleagues in breast cancer, in painstakingly exploring the exact risks and benefits of different adjuvant treatments in different subgroups over the coming years. As an early indicator of some of the factors that may have to be considered, immunohistochemical profiling in the IALT study of the nucleotide excision repair gene product ERCC1—associated with platinum resistance through adduct repair—demonstrated that only ERCC1-negative tumors had significantly prolonged survival compared to observation (HR = 0.67, 95% CI = 0.51-0.89).
Neoadjuvant Chemotherapy and Chemoradiotherapy
A number of theoretical advantages, as well as disadvantages, have become apparent for neoadjuvant chemotherapy vs adjuvant chemotherapy. In the early to mid-1990s, two small randomized trials of neoadjuvant cisplatin-based chemotherapy in stage IIIA disease were reported that changed many US oncologists' practice, at least for a time.
Both studies suggested that survival in stage IIIA disease (using the 1986 AJCC staging system, hence including some cases of T3, N0, M0) could be dramatically improved by treatment plans containing elements of neoadjuvant chemotherapy.[24,25] Although several comparably sized or larger studies across stages I-IIIA have subsequently shown similar benefits in terms of hazard ratios, none have proven statistically significant to date.[26-28] When the original Roth study was reanalyzed at 82 months' follow-up, the actual (no longer estimated) median and 3-year survivals were also no longer significantly different between the two arms. In contrast, when the Rosell study was reanalyzed at 7 years' follow-up, the median survival was still 22 vs 10 months (P = .005). The results of several large neoadjuvant chemotherapy trials, using more modern drug regimens and covering all stages of disease from IA to IIIA are awaited.
In 2003, the Intergroup 0139 study addressed the issue of trimodality neoadjuvant chemoradiotherapy vs bimodality definitive chemoradiotherapy in potentially resectable N2, IIIA disease (T1-3, N2, M0). The investigators found no difference in overall survival but an excess of early deaths in the surgical arm, mostly from acute respiratory distress syndrome and mostly in patients undergoing right pneumonectomies. Median progression-free survival favored the surgical arm. While some surgeons may interpret this as showing a potential benefit for surgery, particularly if high-risk right pneumonectomies could be avoided, the trimodality approach currently cannot be recommended as standard. The only notable exception is in the treatment of T3-4, N0-1, M0 superior sulcus (Pancoast) tumors, for which trimodality therapy is currently accepted as the standard of care.
Pragmatic, although not always data-driven, algorithms and case-by-case decision-making for the management of most stage IIIA and some IIIB disease are still needed: ie, neoadjuvant chemotherapy for all, chemotherapy or chemoradiotherapy only for those with anticipated difficult resections, and so forth. Given the large numbers required to show benefit in the adjuvant setting and the suggestion of risk-benefit ratios specific to molecular biology, disease stage, and possibly drug therapy, more randomized studies will be needed to truly tell us whether (or in whom) neoadjuvant chemotherapy or chemoradiotherapy, with or without targeted therapy, adds anything beyond what can be achieved with modern adjuvant regimens.