Current Concepts in the Diagnosis and Management of Small-Cell Lung Cancer
Current Concepts in the Diagnosis and Management of Small-Cell Lung Cancer
Despite a decreasing incidence in the United States, small-cell lung cancer (SCLC) remains a major clinical problem, with approximately 30,000 new cases each year. The diagnosis of SCLC is usually not difficult. The Veterans Administration Lung Study Group (VALSG) staging system is less accurate than the American Joint Committee of Cancer tumor-node-metastasis (TNM) system (7th edition) at predicting survival in SCLC, especially in lower stage disease. Surgery has not played a major part in the management of SCLC, but emerging data suggest that resection may have a role in earlier stage disease. While the frontline treatment of SCLC has not changed significantly in the past decade, newer agents that are currently being investigated provide hope for better treatment of relapsed/refractory disease for the future.
Small-cell lung cancer (SCLC) has a behavior distinct from its more common counterpart, non–small-cell lung cancer (NSCLC). The incidence of SCLC is apparently decreasing in the United States. In 1993, SCLC comprised approximately 25% of all lung cancer, but a recent Surveillance Epidemiology and End Results (SEER) database analysis found that since then the incidence had decreased to approximately 13%. The explanation for this decline in incidence is likely multifactorial. The disease occurs almost exclusively in smokers, and smoking habits have changed in the past few decades. Cigarettes have also changed, with filter cigarettes gaining considerably in popularity. Another, less well-recognized cause may be the changes in the World Health Organization (WHO) and International Association for the Study of Lung Cancer (IASLC) classifications of lung and pleural tumors that were published in 1999. These revisions have made the criteria for a diagnosis of SCLC more restrictive. Nonetheless, SCLC continues to be a major clinical problem, with an aggressive clinical course and short disease-free duration after initial therapy. This review will focus on the diagnosis, staging, and role of surgery, as well as treatment in relapsed SCLC.
SCLC was considered to be either a form of lymphoid neoplasm or an unusual sarcoma until 1926, when Barnard characterized the tumor and called it “oat cell carcinoma.” The disease is diagnosed with light microscopic assessment. According to the current 2004 WHO definition, SCLC is a “malignant epithelial tumour consisting of small cells with scant cytoplasm, ill-defined cell borders, finely granular nuclear chromatin, and absent or inconspicuous nucleoli. The cells are round, oval and spindle-shaped. Nuclear molding is prominent. Necrosis is typically extensive and the mitotic count is high.” The WHO classification includes only one variant—combined small cell carcinoma—an SCLC with an admixed non–small-cell component (adenocarcinoma, squamous cell carcinoma, large cell carcinoma, or spindle cell or giant cell carcinoma). Although multiple synonyms (anaplastic small-cell carcinoma, small-cell undifferentiated carcinoma, small-cell neuroendocrine carcinoma, oat cell carcinoma, mixed small-cell/large-cell carcinoma) preceded the current terminology and classification, use of these terms is discouraged to avoid confusion.
Despite the appearance of the words “small cell” in the tumor’s name, size alone does not reliably separate SCLC from other lung tumors, including other neuroendocrine lung tumors. In addition to cell size (2 to 3 times the diameter of a resting lymphocyte), the neoplastic cells should have the proper array of characteristic nuclear features, which often include nuclear molding, and a high mitotic rate.[8,9] While there are exceptions, a diagnosis of SCLC is unlikely if the presentation is an isolated peripheral lung lesion without a central mass or if the patient is a lifelong nonsmoker.
Since most SCLCs are metastatic at the time of presentation, in 90% of cases the primary diagnosis is made on the basis of a small biopsy and/or a cytologic specimen (Figure 1). With both approaches, there are quantitative and qualitative limitations in the resulting diagnosis that should be recognized. The tumor sample often has significant necrosis and “crush” artifact that limit the amount of intact tumor available for assessment. The crush artifact, although a frequent occurrence in SCLC, is nonspecific and may also be seen in a variety of benign and malignant lesions. Ancillary studies performed on suboptimal specimens without morphologically intact cells are fraught with error. If the tumor does not qualify as a small-cell carcinoma on the basis of morphologic assessment of an appropriately stained slide, ancillary studies are unlikely to accurately establish the diagnosis.
SCLC is the most common lung tumor in the spectrum of pulmonary neuroendocrine malignancies; the latter include typical carcinoid (TC), atypical carcinoid (AC), large-cell neuroendocrine carcinoma (LCNEC), and small-cell carcinoma (SCLC).[7,8] TC is at the low-grade end of the spectrum, while LCNEC and SCLC are biologically aggressive, high-grade neuroendocrine lung tumors. Five-year survival rates vary from 5% or less for SCLC to 95% for TC. SCLC is distinguished from TC and AC by the morphologic findings described above and by the substantial differences in mitotic activity and necrosis. TC has a low mitotic count (less than 2 mitoses/2 mm2), and AC has an intermediate count (2 to 10 mitoses/2 mm2) with punctate necrosis. SCLC is a high-grade tumor with ample mitotic activity (more than 10 mitoses/2 mm2), averaging 60 to 80 mitoses/2 mm2, and often with widespread necrosis. Within the pulmonary neuroendocrine tumor group, the distinction of SCLC from LCNEC is the most problematic, and the highest reproducibility in diagnoses is found in the distinction of SCLC from TC.
The reason for the difficulty in distinguishing SCLC from LCNEC is that the distinction is primarily based on morphologic assessment of routine stains/preparations. The morphologic differences involve differences in cell size and, more importantly, nuclear characteristics. LCNECs are uncommon (1% to 3% of lung cancers worldwide), and 85% are large peripheral lesions, in contrast to SCLCs, which are centrally located in 95% or more of cases and are substantially more common (15% to 20% of lung cancers worldwide). Other entities in the differential of a mitotically active “small-cell” lesion include lymphoid lesions (benign or malignant), other lung cancers (squamous, basaloid, blastoma), metastatic lesions, and “small blue-cell tumors” (Ewing sarcoma/primitive neuroectodermal tumor [PNET)] Merkel cell carcinoma, rhabdomyosarcoma, neuroblastoma).
Of the immunoperoxidase antibody panels used in the assessment of SCLC, CD45/Leucocyte Common Antigen (LCA) and a keratin cocktail are perhaps the most useful. In a high-grade “small-cell” tumor with the proper morphologic features on light microscopy, a positive result on keratin staining (often demonstrating a perinuclear punctate pattern) and a negative result on testing for CD45/LCA support the diagnosis of SCLC and exclude lymphoid lesions. Neuroendocrine markers (chromogranin, synaptophysin, CD56/Neural Cell Adhesion Molecule [NCAM]) (Figure 1), and thyroid transcription factor-1 (TTF-1) are also commonly used in potential cases of SCLC. With the use of broad spectrum keratin cocktails, SCLCs yield positive results in 90% to 100% of cases and are positive for TTF-1 in 70% to 90% of cases.[8,10] Tests for individual neuroendocrine markers are positive in 50% to 60% of cases but results, even for a panel of markers, may be negative in a significant minority of cases (10% to 30%). Positive results on tests for neuroendocrine markers are not specific for SCLC. Positive reactions may be seen in up to 30% of NSCLCs (including squamous cell carcinomas, adenocarcinomas, and some large-cell carcinomas). Likewise, testing for TTF-1 may be positive in small-cell carcinomas of non-pulmonary origin. In many cases with high-quality standard hematoxylin and eosin sections and well preserved tumor cells, a reliable diagnosis of SCLC can be made using routine preparations without immuno-morphologic studies.
Even in the hands of experienced lung pathologists, around 5% of cases of SCLC have morphologic and immuno-marker characteristics that are difficult to interpret. Studies have repeatedly demonstrated a continuum between SCLC and other non–small-cell carcinomas with respect to morphometrically assessed tumor cell size and nuclear area. That morphologic continuum, in combination with cytologic heterogeneity within the same tumor, limited tumor sampling, and fixation/processing/staining artifacts can make the diagnosis of SCLC challenging. Fortunately, with current methods the pathologic diagnosis of SCLC is reproducible; both SEER data and a prior study of Southwestern Cancer Study Group data suggest a greater than 90% agreement between the submitting pathologist and central review pathologists in the diagnosis of SCLC.[14,15]
Two systems are currently used to stage SCLC: the tumor-node-metastasis (TNM) classification, and the Veterans Administration Lung Study Group (VALSG) limited disease–extensive disease (LD vs ED) system. For decades, the VALSG system was the basis for treatment recommendations. According to this classification system, limited-stage disease is disease that is confined to the ipsilateral hemithorax and within a single radiation port (TNM stages I through IIIB), while extensive-stage disease includes metastatic disease outside the ipsilateral hemithorax. This schema is problematic since patients with extremely limited disease are grouped with those who have locoregionally advanced disease and a worse prognosis. Patients with features of more locally advanced disease (eg, supraclavicular lymph nodes, isolated pleural effusion) are frequently excluded from protocols for limited stage disease, although the prognostic impact of these findings remains controversial.[17,19]
Recently, the IASLC conducted an analysis of 12,620 eligible cases of small-cell histology. TNM staging was available for 8088 of the patients. Survival was directly correlated with both T and N status. Differences were more pronounced in patients without mediastinal or supraclavicular nodal involvement. The prognosis in patients with pleural effusion, regardless of cytology, was intermediate—between those for limited and extensive disease. In another analysis of 349 patients from the same IASLC database who had undergone surgical staging for SCLC, survival after resection correlated with both T and N status. The nodal status had a stronger influence on survival.
Current treatment recommendations are based on the older staging system, but in view of the better correlation of survival with TNM staging, the IASLC currently recommends that the TNM staging system be adopted more uniformly. Future clinical trials for SCLC should also incorporate the staging system described in the seventh edition of the American Joint Committee on Cancer (AJCC) staging classification.[21,22]