INTRODUCTION — Small cell lung cancer (SCLC) is distinguished from non-small cell lung cancer (NSCLC) by its rapid doubling time, high growth fraction, and the early development of widespread metastases. Although considered highly responsive to chemotherapy and radiotherapy, SCLC usually relapses and becomes refractory to treatment within one to two years.

The epidemiology, pathology, clinical presentation, staging, and prognostic factors of SCLC will be reviewed here. The treatment of SCLC is discussed separately.

EPIDEMIOLOGY — SCLC occurs almost exclusively in smokers and appears to be most common in heavy smokers

The proportion of lung cancer in the United States that is classified as SCLC has steadily decreased. This was illustrated by an analysis of the Surveillance, Epidemiology and End Results (SEER) database, in which the proportion of SCLC declined from 17 percent in 1986 to 13 percent in 2002 . This overall decrease in incidence was accompanied by an increase in the percentage of cases of SCLC arising in women (28 percent in 1973 versus 50 percent in 2002).

These epidemiologic changes may be due to differences in smoking patterns (eg, decreased in men) [2]. In addition, changes in the pathologic criteria for SCLC may have led to a classification of some cases as having large cell neuroendocrine carcinoma

PATHOLOGY

Classification of SCLC — The histologic classification of SCLC has evolved over several decades. Initially, the World Health Organization (WHO) classified SCLC into the following three histologic subtypes:  oat cell, intermediate cell type, and a combined oat cell type (SCLC combined with squamous cell or adenocarcinoma). However, the classic oat cell and intermediate subtypes did not appear to have different biologic or clinical behavior patterns. As a consequence, a later proposal by the International Association for the Study of Lung Cancer (IASLC), classified SCLC into pure SCLC, and two less common variants, mixed small cell and large cell (accounting for 3 to 6 percent of cases), and combined small cell type (1 to 3 percent)

The most recent 1999 combined WHO/IASLC classification of lung and pleural tumors eliminated intermediate cell carcinoma, and added a new category of large cell neuroendocrine carcinoma [4,5]. Thus, the three categories of neuroendocrine lung tumors are as follows:

• Classical small cell carcinoma (SCLC)
• Large cell neuroendocrine cancer
• A combined small cell carcinoma, consisting predominantly of SCLC with some areas of NSCLC

The similarity in biologic behavior between small cell and large cell neuroendocrine tumors is supported by their characteristic gene expression profiles

A diagnosis of SCLC is made primarily on light microscopy. Small cell carcinoma is characterized by small "blue" malignant cells about twice the size of lymphocytes. The cytoplasm is sparse, and nuclear features include finely dispersed chromatin without distinct nucleoli. In contrast, the cells of large cell neuroendocrine carcinoma are distinguished by the presence of prominent nucleoli, a more granular chromatin pattern, and variable amounts of cytoplasm. Mitotic rates are high, and necrosis of individual tumor cells is common. Nuclear molding is considered characteristic in well-preserved specimens, although a nondiagnostic "crush" artifact is more frequently observed. The neoplastic cells are typically arranged in clusters, sheets, or trabeculae, separated by a delicate fibrovascular stroma.

In the combined cell subtype, SCLC more often coexists with squamous cell carcinoma rather than adenocarcinoma. While coexistence with another cell type of carcinoma is rarely detected in untreated specimens, up to 30 percent of autopsies with SCLC demonstrate areas of differentiation into non-small cell carcinoma. This finding has led to the hypothesis that pulmonary carcinogenesis occurs in a pluripotent stem cell capable of differentiation along several pathways.

The prognostic significance of mixed large cell and small cell, or combined small cell with squamous cell or adenocarcinoma is unclear. For the mixed small cell and large cell subtype, published series indicate survival that is inferior to, superior to, or comparable to pure SCLC. There is also little available information regarding the prognostic impact of combined SCLC/NSCLC histology. In some reports, they are more resistant to treatment and have a short survival . However, others suggest a more favorable prognosis. In one series of 429 patients with SCLC treated at Vanderbilt University, two of nine patients with combined small cell and non-small cell histology were long-term survivors, and both underwent surgical resection in addition to chemotherapy, suggesting a potential role for surgery in this subgroup

Tumor markers — Eight SCLC antigen clusters have been identified by segregation analysis, and divided into three groups: neural, epithelial, and neuroendocrine

• Because of their epithelial cellular origin, virtually all SCLCs are immunoreactive for keratin and epithelial membrane antigen.

• Neuroendocrine and neural differentiation result in the expression of dopa decarboxylase, calcitonin, neuron-specific enolase, chromogranin A, CD56 (neural cell adhesion molecule [NCAM], gastrin releasing peptide (GRP), and insulin-like growth factor-I (IGF-I). One or more markers of neuroendocrine differentiation can be found in approximately 75 percent of SCLCs. In contrast to the large cell neuroendocrine carcinoma, the immunohistochemical demonstration of neuroendocrine differentiation is not a prerequisite for the diagnosis of small cell carcinoma [5].

Occasional patients can produce autoantibodies that crossreact with both SCLC cells and the central nervous system or the neuromuscular junction. These autoantibodies can cause cerebellar degenerative syndromes or the Lambert Eaton myasthenic syndrome. SCLC is the most common malignancy associated with neurologic paraneoplastic syndromes.

SCLC cells can also produce a number of polypeptide hormones, including ACTH and vasopressin (antidiuretic hormone), resulting in various paraneoplastic and ectopic hormonal syndromes

IGF-I and GRP may also participate in autocrine and paracrine loops to enhance the growth of SCLC cells. In addition to overexpression of these growth factors, the enzyme which inactivates these small polypeptide hormones, neural endopeptidase (CALLA), is underexpressed in SCLC cells, and in bronchoalveolar lavage (BAL) fluid from smokers

Genetic abnormalities — The development of both SCLC and NSCLC occurs through stimulation of proliferation and mutagenesis occurring over years, and resulting from exposure to tobacco and other carcinogens. Multiple genetic defects have been detected; some are characteristic and perhaps involved with oncogenesis, whereas others are probably random or secondary events. A detailed discussion of the molecular genetic changes associated with small cell neuroendocrine tumors is beyond the scope of this review. The most common characteristics of SCLC are summarized below

• p53 mutations are detected in 75 to 90 percent of SCLCs

• Loss of heterozygosity of chromosomes 9p and 10q (the site of the PTEN gene) is present in the majority of tumors

• Deletion of 3p, including 3p21-22, leading to inactivation of as many as three putative tumor-suppressor genes. The FHIT gene (for fragile histidine triad), which is involved in the accumulation of diadenosine tetraphosphate, thereby leading to DNA synthesis and proliferation, has been localized to 3p14.2 and is believed to be an important tumor suppressor gene involved in the pathogenesis of lung cancer]. A second postulated tumor suppressor gene located at 3p21.3 is RASSF1A, which is absent in all SCLCs tested. Deletion of a third gene, TGFBR2, located at 3p21.3.22, and encoding the transforming growth factor beta type II receptor, has also been described in SCLC

• Loss of the retinoblastoma gene function at 13q14 is nearly ubiquitous in SCLC]. As an example, approximately 60 percent of SCLC cell lines have undetectable transcripts, while the remaining 40 percent have an abnormal gene product.

• Telomerase is a ribonucleoprotein enzyme that compensates for telomere shortening during cell division by synthesizing telomeric DNA, thereby maintaining telomere length. In normal somatic cells, telomerase activity is usually undetectable, with the exception of some cell types that possess the ability to divide indefinitely (eg, hematopoietic cells, hair follicles, intestinal crypt cells, and basal cells of the epidermis). Activation of telomerase is detected in approximately 90 percent of SCLCs. In cancer cells, telomerase activity correlates with the stabilization of telomere length and cellular immortalization.

• Upregulation of wild type c-kit and its phosphorylated form may represent another autocrine/paracrine growth factor loop and is detected in up to 80 to 90 percent of SCLCs, although imatinib has not demonstrated activity in phase II trials.

• Through gene expression arrays and CCGH analyses, upregulation of the c-myc pathways, the islet-1 and the forkhead box protein G1B transcription factors appear to be characteristically upregulated in SCLCs

• In contrast to NSCLC, mutations in the K-ras oncogene and p16 abnormalities are rare.

CLINICAL PRESENTATION — SCLC typically arise in the central airways, infiltrating the submucosa, and it gradually obstructs the bronchial lumen through extrinsic or endobronchial spread. The most common presentation is that of a large hilar mass with bulky mediastinal adenopathy. Potential clinical consequences include cough, dyspnea, weight loss, and debility, with or without postobstructive pneumonia. Approximately 70 percent of patients present with overt metastatic disease; SCLC has a particular propensity to spread to liver, adrenals, bone, bone marrow, and brain.

SCLC can occasionally present as a peripheral nodule. However, a solitary peripheral nodule without central adenopathy is sufficiently uncommon that cytologic diagnosis from a fine needle aspirate might not adequately differentiate SCLC from typical or atypical carcinoid, or from a well-differentiated neuroendocrine tumor. In this situation, mediastinal staging followed by surgical resection is recommended.

Finally, small cell lung cancer can occasionally present with a paraneoplastic syndrome.

STAGING — The two stage system originally introduced by the Veterans' Affairs Lung Study Group (VALSG) is widely utilized in staging of SCLC because of its simplicity and clinical utility:

• Limited disease is defined as disease confined to the ipsilateral hemithorax and within a single radiotherapy port (corresponding in part to TNM stages I through IIIB,

• Extensive disease is defined as evident metastatic disease outside the ipsilateral hemithorax.

At presentation, 60 to 70 percent of patients with SCLC have extensive stage, and 30 to 40 percent have limited stage disease. This distinction is clinically relevant, since patients with limited stage disease are generally treated with combined modality therapy, while those with extensive stage disease receive chemotherapy alone.

This schema is problematic when applied to patients with extremely limited disease or with locoregionally advanced disease, such as those with contralateral hilar or supraclavicular nodes, pericardial effusions, or malignant pleural effusions. Patients with features of more locally advanced disease are frequently excluded from protocols for limited stage disease, although the prognostic impact of these findings remains controversial. Not surprisingly, the extent of locoregional disease appears to correlate with survival in patients with limited stage disease who undergo combined modality therapy

TNM or IASLC staging system — Although the more descriptive TNM staging system for NSCLC that is advocated by IASLC) describes the extent and burden of disease more accurately than the VALSG schema, it is of limited value in most cases of SCLC because of an overlap in prognosis and therapy between stage groupings, particularly stage II and III . It is useful for patients with SCLC that is limited to the ipsilateral lung (stage I disease), in whom resection followed by adjuvant chemotherapy may be feasible; however, this applies to fewer than 10 percent of patients.

Of note, the definition of "limited disease" in the TNM/IASLC staging system includes any patient without distant metastatic disease. Although not in widespread use, this definition may provide better prognostic stratification than the VALSG definition of limited versus extensive disease. In one report that compared outcomes of patients with limited stage disease by VALSG or IASLC criteria, the prognosis of patients with limited disease by IASLC but not VALSG criteria was similar to that of others with limited stage disease by both criteria .

Staging workup — Systemic therapy is required for all patients with SCLC, even those with radiographically limited stage disease. Therefore, the major therapeutic significance of staging is to guide the use of chest radiotherapy, which is indicated for limited but not necessarily for extensive disease. After a tissue diagnosis is made, an exhaustive staging workup includes physical examination, chest radiography, chest, liver, and adrenal computed tomography, head CT or cranial MRI (particularly for asymptomatic patients, bone scan, and, in selected cases, unilateral or bilateral bone marrow aspirates and biopsies. Bone marrow is involved in 15 to 30 percent of patients at presentation, but it represents a solitary site of metastatic disease in only 2 to 6 percent of cases

Outside of the context of a clinical trial, an alternate, abbreviated staging algorithm could be directed by symptoms, and terminated as soon as extensive disease is documented. This latter strategy, the rule for many European trials, is likely to become more prevalent in the United States in this time of health care parsimony and need for economic justification. Specifically, given the relative invasiveness of bone marrow examination, many investigators argue that this test has a very low yield in patients with normal serum lactate dehydrogenase (LDH) levels

Arguments against sequential staging include the fact that bone scans are positive in up to 30 percent of patients without symptoms or an abnormal alkaline phosphatase. Furthermore, head CT scan is positive in about 15 percent of patients at diagnosis, including 5 to 8 percent of asymptomatic patients . Early identification and treatment of brain metastases is important, as it results in a lower rate of chronic neurologic morbidity.

Of greater importance than the complete ascertainment of every involved site of disease is the rapid tempo of untreated SCLC. Most patients develop their prediagnosis symptoms within eight weeks. Staging should not delay treatment onset more than 7 to 10 days, as many patients will become seriously ill in the interval.

Utility of PET scanning — Positron emission tomography (PET) with 18-fluoro-2-deoxyglucose (FDG) is a noninvasive method to assess for mediastinal or distant metastases by detecting areas of high FDG uptake (ie, increased metabolic activity) consistent with malignancy. Although PET has been widely studied as a component of the staging evaluation for non-small cell lung cancers, there are few published studies in patients with SCLC

The utility of PET for staging SCLC was prospectively studied in 24 patients who were assessed by conventional means as having limited stage disease. The PET scan was positive at all sites of previously documented disease involvement in all patients (sensitivity 100 percent). In addition, three patients had PET scan findings suspicious for extensive stage disease, although only two reflected true disease spread that altered the treatment plan (one with bone metastases and a negative bone scan, and one with contralateral supraclavicular nodal uptake). The third patient, who had contralateral lung FDG uptake, was diagnosed as having a fungal infection based on high-resolution CT.

PROGNOSTIC FACTORS — From the time of diagnosis, the median ranges of survival for limited and extensive disease are 15 to 20 months and 8 to 13 months, respectively. Approximately 20 to 40 percent of limited stage and less than 5 percent of extensive stage patients survive two years. The respective values for five-year survival are 10 to 13, and 1 to 2 percent, respectively. The most important adverse prognostic factors are:

• The host factors of poor performance status and weight loss

• Tumor-related factors such as the extent of disease (limited versus extensive). In limited disease, early stage disease (stage I) carries a favorable prognosis, while an elevated LDH is unfavorable. In extensive disease, the number of organ sites involved is inversely related to prognosis. Metastatic involvement of the central nervous system, the marrow, or the liver is unfavorable compared to other sites, although these variables are confounded by the number of sites of involvement.

• In most trials, women fare better than men, although the reasons for this are not known.

• The presence of paraneoplastic syndromes is generally unfavorable.

With longer duration of follow-up, the prognostic importance of many of these variables diminishes, largely as a result of the smaller numbers of long-term survivors

SUMMARY AND RECOMMENDATIONS

• Small cell lung cancer (SCLC) is distinguished from non-small cell lung cancer (NSCLC) by its rapid doubling time, high growth fraction, and the early development of widespread metastases. Although SCLC is highly responsive to chemotherapy and radiotherapy, disease usually relapses within two years despite treatment. Overall, only three to eight percent of all patients with SCLC survive beyond five years.

• SCLC is one of three types of neuroendocrine lung cancer. The others are large cell neuroendocrine cancer and combined small cell carcinoma, which consists predominantly of SCLC with some areas of NSCLC. Whether these histological types have different prognoses is uncertain.

• SCLC is generally diagnosed by light microscopy. It is characterized by small "blue" malignant cells about twice the size of lymphocytes. The cytoplasm is sparse, and nuclear features include finely dispersed chromatin without distinct nucleoli.

• Eight antigen clusters have been identified and multiple genetic defects have been detected in SCLC, especially deletion of 3p, including 3p21-22, and loss of RB1 function.

• SCLC typically arise in central airways of smokers, and due to their biological aggressiveness, have a brief period of pre-diagnosis symptoms, yet present with overt metastatic disease in two-thirds of patients.

• SCLC is the most common malignancy associated with neurologic paraneoplastic syndromes and is frequently associated with ectopic hormonal syndromes such as SIADH and Cushing's Syndrome.

• Although chemotherapy remains the mainstay of therapy, chest radiotherapy given concurrently with chemotherapy increases the likelihood of long-term (two and five year) survival for patients with limited stage disease. Thus, extensive staging is indicated to detect whether overt metastases are present.