Bronchial carcinoma occurs more frequently in the right lung than in the left in a ratio of approximately 6 to 4. The upper lobes are involved more often than the lower lobes, and the middle lobe is involved the least frequently. In the upper lobes, the tumor is most likely located in the anterior segment, although the other segments are not spared as the site of origin.

The anatomic site of origin of the tumor may be classified as (a) the central zone, including the main stem, lobar bronchi, and primary segmental bronchi of the lower lobe; (b) the segmental or intermediate zone, including the third-, fourth-, and possibly the fifth-order segmental bronchi; and (c) the peripheral zone, which includes the remainder of the distal bronchi, bronchioles, and alveoli. In the radiographic localization of lung tumors, zones 1 and 2 may be considered the central area and zone 3 the peripheral area.

According to Meyer and Liebow,¹⁶¹ approximately 50% to 60% of carcinomas of the lung originate in the peripheral area. Of those carcinomas that arise in the central and segmental zones (the central area), 20% to 40% arise in the former and 60% to 80% in the latter.

Grossly, the tumors in the central and segmental zones appear as firm, irregular masses of varying size. Intraluminal growth may occlude the bronchial lumen partially or completely, although obstruction may also be caused by circumferential narrowing of the lumen. Extrabronchial spread may extend for a variable distance into the adjacent lung parenchyma. The tumor is generally homogeneous, with a whitish gray cut surface. The endobronchial surface is typically ulcerated. Atelectasis or secondary inflammatory change—including secondary bronchiectasis, pneumonitis, or lung abscess distal to the site of the tumor—is frequently present.

The peripherally located tumors are firm and irregular and may or may not appear to be demarcated from the surrounding lung tissue. The cut surface is homogeneous. The smaller lesions are usually solid, although the larger ones may reveal central necrosis with cavitation. Umbilication or puckering of the overlying adjacent visceral pleura is often present. The blood supply of both the centrally and peripherally located bronchial carcinomas is from the bronchial arteries.

Many histologic classifications of lung tumors have been suggested. A modified classification from the original suggested by Moori¹⁷³ is shown in Table 104-1. This classification is less cumbersome than that proposed by the World Health Organization and International Association for the Study of Lung Cancer published in 1999 by Travis and associates.²⁷¹ In clinical practice it is common to separate these tumor types into either small-cell cancer (SCLC) or non-small-cell cancer (NSCLC). Furthermore, the SCLCs belong to the spectrum of neuroendocrine tumors of the lung. Azzopardi¹⁴ and Bensch and associates²⁰ were among the early investigators who established the recognition of the neuroendocrine features of these tumors. Subsequently, the initial classification has undergone numerous modifications, with better delineation and characterization of the various subtypes in the neuroendocrine tumor classification (Table 104-2). Arrigoni and colleagues⁸ identified the atypical carcinoid tumors. Gould and coworkers,^79,80 incorporating the increased knowledge concerning these tumors, suggested that the atypical carcinoid tumors be termed well-differentiated neuroendocrine carcinomas. They further suggested the term neuroendocrine carcinoma of the intermediate cell type be applied to all types of poorly differentiated neuroendocrine tumors that were distinct from either the atypical carcinoids or the small-cell type of neuroendocrine tumors. Travis and associates²⁷² proposed a new category of these tumors: the large-cell neuroendocrine tumors. The latter authors and colleagues^273,274 have further defined the criteria for the classification of the large-cell neuroendocrine tumors. Wick²⁹³ suggests that a new terminology would be better for the neuroendocrine lung tumors: (a) grade I, well-differentiated neuroendocrine tumor (carcinoid); (b) grade II, moderately differentiated neuroendocrine tumor (atypical carcinoid); and (c) grade III, poorly differentiated neuroendocrine tumor (i) of the large-cell type and (ii) of the small-cell type.

Grossly, in situ carcinoma appears as an area characterized by a loss of the normal longitudinal folds of the mucosa of a major or even a segmental bronchus. The mucosal area involved may be thickened and erythematous. Microscopically, the in situ carcinoma reveals full-thickness cytologic atypia with an increased nuclear/cytoplasmic ratio. The nuclei are hyperchromatic, and mitoses may be present. These changes do not extend beyond the basement membrane. Often such changes are associated with frank squamous cell carcinoma, but—as observed by Auer and associates¹³—complete regression of an isolated in situ lesion may occur.

Atypical adenomatous hyperplasia (AAH) is a proliferation of bronchioloalveolar cells that is thought to be a precursor to adenocarcinoma. According to Ritter,²³⁸ AAH resembles the nonmucinous variant of bronchioloalveolar carcinoma (BAC). Initially, these lesions were identified pathologically in lungs resected for adenocarcinoma. Rao and Fraire²²³ noted that these
areas of hyperplasia are present in up to 20% of resected lung cancer specimens as an ill-defined peripheral nodule, most often multiple in number. The size of the lesions, according to Weng and colleagues,²⁹¹ was noted to range from 1 to 10 mm, with the majority being ≤5 mm in size. With the advent of helical computed tomography (HCT) for lung cancer screening, many more small lesions <10 mm in size are being identified. Such lesions may be further analyzed with the use of high-resolution CT (HRCT), as reported by Reeves and Kostis.^225,226 The resulting images contain a great deal more information, and the size and shape of the lesion can be precisely measured and recorded. The opacity of the AAH lesion is indeterminate but at times may be “ground glass” in appearance.

Histologically, the lesion appears as a zone of alveolar wall thickening as the result of the alveoli being lined by a variable number of cuboidal or columnar cells (Fig. 104-1A,B). Kerr¹²¹ notes that many of the cells have a hobnail or pear shape, reminiscent of Clara cells, while others show nuclear inclusions as seen in type II pneumocytes. There is variable nuclear pleomorphism, but it is most often mild in nature. There is a lack of any central fibrosis, and mucin production is absent. Kitamura and associates¹²⁵ have classified these atypical adenomatous hyperplasias as low-grade, high-grade, and carcinoma-like. The proliferative potential and p53 expression increase as the grade advances. Likewise, Kitamura and colleagues¹²⁶ recorded that the nuclear area and lesion size increased as the dysplastic lesion progressed to a bronchioloalveolar carcinoma; also, an increase in the level of carcinoembryonic antigen (CEA) was seen as these changes occurred. Any lesion >5 mm should be considered as a probable carcinomatous tumor. Vazquez and Flieder²⁷⁹ have stressed that the distinction between atypical adenomatous hyperplasia and bronchioloalveolar carcinoma is extremely difficult and subjective. The interpretation between a benign lesion and a carcinoma by examination of a fine-needle aspiration of such lesions is not possible in the absence of finding a diagnostic malignant cell. If no malignant cell or cells are identified, a tentative diagnosis of atypical adenomatous hyperplasia may be made. However, the lesion should be reevaluated by HRCT in 4 to 6 months to determine whether it is stable or if growth has occurred. If growth is apparent, the degree and rate of growth can be established by the techniques described by Yankelevitz and associates.^297,298 Growth implies the presence of cancer, therefore further diagnostic evaluation or removal of the lesion is indicated.

As a consequence of an increased use of routine helical examination of the lungs and evaluation of any lesion identified by HRCT, a subset of small lesions has been identified as being solitary pure ground-glass opacities. Any areas of fibrosis or scarring exclude the lesion from this category of indeterminate masses. The natural history of these pure ground-glass opacities has heretofore been unknown. In an attempt to evaluate the natural history and characteristics of these opacities, Kodama and associates¹²⁷ investigated and reported 49 patients with a solitary pure ground-glass opacity in the lung discovered by HRCT. Of these individuals, 39 had undergone surgical resection within 19 months of the initial identification of the lesion; 34 had BAC, 4 had adenocarcinoma, and 1 had AAH. The average diameter of all lesions was 13.3 mm, but in the 4 patients with adenocarcinoma, the average diameter was 27 mm. Nineteen additional patients with pure ground-glass opacities (17 patients with solitary lesions, 1 patient with two lesions, and 1 with many lesions) were observed for a period of 2 years or longer. Reasons for observation varied, but eventually 10 of the patients agreed to have their lesion removed; nine others elected continued observation. Overall, in the 19 patients the lesion increased >5 mm in size in 5 patients, increased slightly in 6, and showed no change in 8. It is of interest that six of the patients who eventually elected to have a resection had a history of a previously resected cancer (five in the lung and one in the kidney). Of the 10 lesions resected, the largest (25 mm) was an adenocarcinoma, 4 lesions were BAC, 1 was AAH, 3 showed the presence of a lymphoproliferative disorder, and 1 revealed the presence of pulmonary fibrosis. Of note is that four of the five patients with proven cancers had a prior history of a previously resected lung cancer. From these data, although not statistically significant, it seems apparent that any pure ground-glass opacity >10 mm in size that shows evidence of growth over a period of observation or is identified in a patient with a history of lung cancer should be strongly considered for resection despite the observation that a prolonged period of observation may not be excessively detrimental. Watanabe and colleagues²⁸⁷ carried out limited resection in 17 patients with pure ground-glass attenuation on HRCT. All were pure BAC
lesions and no recurrence or death occurred on short-term follow-up (median follow-up time was 32 months).

Bronchioloalveolar carcinoma (BAC) represents highly differentiated adenocarcinoma, but many pathologists have considered it a separate and distinct tumor. Nonetheless Moori,¹⁷³ Noguchi,¹⁸⁶ and Kondo¹²⁹ and their associates, among others, believe that these tumors represent a unique subset of adenocarcinoma. They constitute 1.5% to 7.0% of all bronchial carcinomas, with an average incidence of 2.5%. However, Barsky and colleagues¹⁶ have noted an increasing incidence of these tumors; in 1955, the bronchioloalveolar carcinomas accounted for only 5% of resected lung cancer specimens, whereas in 1990, these tumors made up 25% of resected lung cancers.

Singh and associates²⁵⁷ described the criteria for the diagnosis of bronchioloalveolar carcinoma as follows: (a) absence of primary adenocarcinoma elsewhere in the body, (b) absence of a central bronchial adenocarcinoma, (c) a peripheral location, (d) growth using the alveolar septa as supporting structures (lepidic growth), and (e) a characteristic histologic appearance different from that of other lung tumors. Last, aerogenous spread of the tumor within the lung is believed to be important.

Microscopically, the alveolar spaces are lined by malignant cuboidal or nonciliated columnar epithelium in layers or in papillary formation (Fig. 104-5). The alveolar spaces may be filled or even distended with this proliferating epithelium. Occasionally, single cells or clusters containing large multinucleated giant cells may appear lying free. The nuclei are hypochromatic, but mitoses are not common. The cytoplasm is acidophilic and abundant, or the cells may be mucous cells with extensive amounts of mucus filling the alveolar spaces. The lung architecture is most often preserved, although a few specimens show diffuse malignant invasion. Scar formation is seen in approximately half the tumors. Also in approximately half the tumors, mucin-producing cells are observed as the dominant cell type. Clayton⁴² initially separated bronchioloalveolar carcinomas into two cell types: mucinous and nonmucinous (i.e., Clara cells and type II pneumocytes). Rarely, the tumor may consist of a mixture of the mucinous and nonmucinous subtypes. Sakurai and coworkers^243,244 reported that BAC tumors <3 cm in size were nonmucinous in 84%, mucinous in 8%, and combined in 8% of cases, respectively.

Clayton⁴³ has divided the bronchioloalveolar carcinomas into three subgroups. Histologically, the first group comprises tumors made up of high columnar, mucin-producing cells with little nuclear atypia: in Clayton’s series this group made up approximately one-fourth of all bronchioloalveolar tumors. This subtype tends to spread aerogenously, forming satellite tumors. The second subgroup is made up of tumors composed of cuboidal or low columnar cells with a high degree of atypia producing little or no mucin; psammoma bodies may be present. The third subgroup is formed by those tumors that have a central area of necrosis and often termed sclerosing bronchioloalveolar tumors. Barsky and associates¹⁶ reported the incidence of the three subtypes as nonmucinous in 48%, mucinous in 42%, and sclerotic in 10%, respectively. Of possible prognostic significance was that areas of dedifferentiation of BAC into solid areas of moderately or poorly differentiated adenocarcinoma were seen in 10%, 27%, and 42% of the nonmucinous, mucinous, and sclerotic types, respectively.

Clayton⁴² had earlier shown by electron microscopy that the bronchioloalveolar carcinomas may be composed of Clara-type cells, type 2 granular pneumocytes, or mucous cells. Barsky and colleagues¹⁷ also identified these three phenotypes in the bronchioloalveolar tumors and have suggested that these findings, as well as the multifocality of these cancers, are evidence of their multiclonal origin. This conclusion was supported to some extent by the study of p53 mutations in primary and secondary lung cancers by Mitsudomi and associates.¹⁶⁶ However, Holst and coworkers¹⁰⁰ studied eight multifocal cases of bronchioloalveolar carcinoma by using topographic genotyping, which enabled them to define the spectrum of point mutational changes in K-ras-2 and p53 oncogenes. They demonstrated that, although no commonality existed between the two gene mutations, the primary tumor and satellite and intrathoracic metastases showed the same point mutations of either oncogene when present. The authors concluded that their findings support a monoclonal origin of multifocal bronchioloalveolar carcinoma, the multifocal disease being the result of intra-alveolar, lymphatic, and aerosol spread.

Immunohistochemical studies by Lee and colleagues¹⁴³ have shown that antibodies used to detect exocrine features that depend on sugar or glycolipid epitopes immunostain alveolar carcinomas regardless of mucin production. These tumor cells also immunostain for high- and low-molecular-weight keratin, epithelial membrane antigen, CEA, and leu-7.

In patients with peripheral adenocarcinomas <20 mm in size, Kondo and associates¹²⁹ identified 102 (84.2%) BAC tumors (aveolar replacement growth), most of which were either type A or B; less than one-third were type C, the more aggressive variety. No lymph node metastases were noted in any patients of this entire group, although metastasis may occur with type C tumors. The 19 (18.6%) patients with type D, E, or F tumors (true adenocarcinoma—nonalveolar replacement growth) likewise had no metastatic node involvement, but lymphatic vessel invasion was noted in three patients.

Grossly, BAC tumors may occur in one of three forms: solitary, multinodular, and diffuse or pneumonic type. The first is the most common and accounts for slightly less than half to almost two-thirds of the tumors of this subclassification. In the series reported by Hill,⁹⁵ the percentage of a solitary mass or nodule was 43%; in the series reported by Harpole,⁸⁷ Daly,⁵⁴ and Ebright⁶⁴ and their colleagues, the percentages were 59%, 69%, and 64%, respectively. Approximately 66% of the solitary tumors in the series of Daly and associates⁵⁴ were ≤3 cm in diameter. Only a small number of these, however, fit the criterion of a solitary pulmonary nodule (i.e., a well-demarcated lesion <3 cm in diameter completely surrounded by pulmonary parenchyma). In the aforementioned series of 184 bronchioloalveolar carcinomas, only an 11.2% incidence of “solitary pulmonary nodules” was recorded. The vast majority of the small lesions may be ill-defined infiltrates, and an increasing number of these are being identified by HRCT. The larger, diffuse tumors may appear as masses, as infiltrates, or even consolidations of an entire lobe. The multinodular, unilateral, or bilateral lesions make up just between 10% and 20% of the total. Direct extension beyond the lung is uncommon and lymph node metastasis is likewise infrequent and, as noted in most series, do not occur with types A and B tumors. Daly and associates⁵⁴ recorded only a 7.5% incidence of lymph node metastasis in tumors of the C type. Ebright and coworkers⁶⁴ reported a similar incidence (7.2%) in the series from the Memorial Sloan-Kettering Cancer Center. The classification of BAC tumors by Ebright and associates⁶⁴ is different than that of Noguchi and colleagues.¹⁸⁶ The former authors⁶⁴ separate BAC tumors into three groups: (a) pure bronchioloalveolar carcinoma, (b) bronchioloalveolar with local invasion, and (c) adenocarcinoma with bronchioloalveolar carcinoma features. The incidence of each of these in their 100 patients was 47%, 21%, and 32%, respectively. Although their data and conclusions are convincing, only those data relative to types (a) (pure BAC) and (b) (BAC with local invasion) are discussed here. Of these 68 cases, 60% were unilocal lesions, 29.4% were multifocal (uni- or bilateral), and 10.2% were pneumonic in nature. The median survival was not reached at the time of publication for pure BAC tumors, whereas it was 52.2 months for the multifocal lesions (for the bilateral multifocal cases it was 75 months); it was only 18.7 months for the pneumonic type. Only four patients developed lymph node metastases, and distant metastases were seen in six cases (the brain and skeletal system being the more common sites). However, local recurrence (locoregional or a distant metastasis) and the appearance of a new primary were common; in pure BAC patients the incidence was 51.2%; 57.1% local recurrence and 42.8% distant spread of the tumor in the thorax. New primary lesions were noted in 18.7%. For patients with BAC with local invasion, the overall incidence was 50% and these cases were equally divided between recurrent disease and new primaries. Of the seven patients with a pneumonic lesion, two had relapses and three developed new cancers. No lymph node metastasis was identified, although one patient died of a brain metastasis.

Adenosquamous carcinomas are defined as tumors composed of an admixture of squamous cell carcinoma and adenocarcinoma. At least 5% of the tumor should be composed of the minority cell type, according to the criteria of Takamori²⁶³ and Shimizu²⁵³ and their coworkers, although the World Health Organization (1982) gives no reference to the actual ratio. The Japan Lung Cancer Society¹¹⁵ has stated that at least 20% of the tumor should be formed by the less dominant variety; this group also notes that the tumors should be categorized as 20% to 40%, 40% to 60%, and 60% to 80% of either one of the cell types. The overall incidence of adenosquamous carcinoma is between 0.4% and 4.0%; in the surgical series reported by Takamori²⁶³ and Shimizu²⁵³ and their associates, the incidence was 2.6% and 3.4%, respectively. These tumors are more often peripheral rather than central in location. Although Ishida and coworkers¹⁰⁶ reported an almost equal occurrence in men and women, Shimizu and associates,²⁵³ in a much larger series (44 patients versus 11 patients), reported a marked preponderance of men (35:9). Three-fourths of these tumors were >3 cm in size, with a range of 1.5 to 6.0 cm. Microscopically, either component may be well, moderately, or poorly differentiated. The squamous cell component was the dominant tumor in the majority of cases in Shimizu and associates’²⁵³ report.

Ichinose and associates¹⁰² reported that the DNA ploidy patterns of both components showed that, despite the different phenotypes, in 67% of the specimens examined, similar biological characteristics were present. These findings support the theory of Steele and Nettesheim²⁷⁸ that these tumors arise from instability of differentiation. Immunohistochemically, pankeratin stained both the squamous and glandular components. The high-molecular-weight keratin stained mostly the squamous component and not the glandular one. The low-molecular-weight keratin stained the glandular component slightly more often than the squamous component, but in most cases, it
did not stain either component to any great extent. CEA and epithelial membrane antigen stained both elements of most tumors with relatively high frequency.

Blood vessel invasion is common, and lymph node metastasis occurs in over 50% of patients. The latter was present in 61% of the specimens in Takamori and associates’²⁶³ series. The incidence of lymphatic metastasis appears to be greater in patients with a higher percentage of adenocarcinoma present in the tumor (Table 104-5). The cell type of the metastatic disease can vary. Overall, the prognosis is poorer than that of either a squamous cell carcinoma or adenocarcinoma of a similar stage.

The aforementioned impression of the poor prognosis of adenosquamous carcinoma of the lung has been further supported by the observations of Nakagawa and coworkers.¹⁷⁶ The surgical results in 30 patients (2.4%) compared with the results in 1,219 cases of either squamous cell or adenocarcinoma treated during the same period showed a cumulative survival rate of 6.2% at 5 years versus one of 41.5%, respectively. The 5-year survival rate of stage I and II (18 patients) was only 9.7% for those patients with adenosquamous cell carcinoma and 55.4% (771 patients) with either squamous cell or adenocarcinoma of the same stages. Furthermore these authors noted that a greater number of patients with adenosquamous cell carcinoma have a higher stage of disease when it is initially identified. The studies of the aforementioned authors, in addition to those of Fitzgibbons and Kern⁶⁹ as well as Naunheim¹⁸³ and Riquet²³⁰ and their colleagues, support these conclusions. It should be noted, however, that Sridhar and associates²⁵⁹ found that there was no major difference in the survival rate of patients with resected local stage disease, albeit a small percentage (10%) of the patients with adenosquamous cell carcinoma and those with other, similarly staged NSCLCs, especially those with adenocarcinoma or large-cell tumor.

Basaloid tumors of the lung are uncommon lesions. They are considered a variety of the non-small-cell, nonneuroendocrine lung carcinomas. Prior to the initial description of this tumor by Brambilla and associates,²⁸ such tumors were considered to be either a poorly differentiated squamous cell tumor or a poorly differentiated large-cell tumor. In fact, in the World Health Organization (WHO) classification,²⁶⁸ the basaloid tumor is listed in both the squamous cell and large-cell categories. Nonetheless, the basaloid tumor is now considered to be a separate entity.

Grossly these tumors are located in a lobar or segmental bronchus in 85% of the cases; 15% are located more distally. The tumor is usually exophytic but is associated with bronchial wall invasion. Brambilla and associates^27,28,29 have described the histological features of this tumor: small, moderately pleomorphic, cuboidal to fusiform cells with hyperchromatic nuclei, graular dense chromatin, and a scanty amount of cytoplasm. The mitotic index is high. Immunohistochemically, the basaloid carcinomas express low-molecular-weight cytokeratins. One of three neuroendocrine markers (discussed subsequently) may be present in up to 40% of the tumors, as in the study by Kim and colleagues¹²³ of 24 patients with a pure form and 11 with a mixed form of the tumor. Sturm and coworkers²⁶¹ reported that the presence of the thyroid transcription factor 1(TTF-1) excluded the diagnosis of basaloid carcinoma. Immunohistochemical staining for 34 β E12 (low-molecular-weight cyto- keratins) was present in 80% of the basaloid cancers. In Kim and colleagues’¹²³ study, 34 β E12 was also present in 80% of patients and TTF-1 was negative in all. Genetic studies have shown that p53 is stabilized in 85% of the basaloid tumors. These tumors are thought to originate from the basal, pluripotent reserve cells of the bronchial mucosa. The major histologic differential diagnoses are a small-cell carcinoma, as discussed by Foroulis and associates,⁷² and large-cell neuroendocrine carcinoma, as noted by Kim and colleagues.¹²³

Basaloid carcinomas of the lung have been considered to be highly aggressive, and despite resection of even early-stage lesions with or without subsequent irradiation, the 5-year survival rate reported by Moro and associates¹⁷⁴ was only 15%. However, Kim and colleagues¹²³ reported a 5-year survival rate of 57.2% after resection of stage I and II disease. This compared favorably with a 5-year survival of 53.9% in 85 stage I and II patients with resected poorly differentiated squamous cell carcinoma. Whether other series will substantiate these results remains to be seen.

It has become apparent that the size of NSCLCs is of greater importance than has been recognized in the past. This is particularly true of small, peripheral tumors ≤3 cm in size, but it is evident that a size of >3 cm is likewise relevant as to the pathologic behavior of a given tumor.

In an early series of patients with tumors ≤3 cm or less in size and without lymph node involvement Read²²⁴ and associates reported that the 5-year survival was significantly greater in the patients with tumors <2 cm in size as compared with those >2 to 3 cm in size. The reports of Port,²¹⁹ Gajra,⁷⁴ Okada,^195,197 and Birim²³ and their coworkers support the aforementioned significant findings (Table 104-6). Wisnivesky and colleagues,²⁹⁴ using a database^* of over 7,000 patients with stage I disease, confirm these findings (Table 104-7).

Patz²¹⁵ and Takeda²⁶⁴ and their associates, however, found no significant differences in similar groups of patients. In fact in
Takeda and coworkers’²⁶⁴ study, only a tumor size >5 cm was an independent significant prognostic factor. Subsequently Ohata and associates¹⁹⁴ observed, in 86 patients with stage I NSCLC ≥5 cm in size, that 56.8% of the tumors relapsed within 12 months. The overall 5- and 10-year survival rates were 42% and 24.2% respectively. Watanabe²⁹⁰ and Carbone³⁶ and their associates also reported that patients with tumors >5 cm in size had a worse survival rate than those patients with smaller tumors. However, in these latter two studies, tumors <2 cm in size had a significantly better prognosis than those >2 cm to 4 cm. Moreover, in an earlier report, López-Encuentia and colleagues,¹⁴⁸ in a study of 1,020 patients with NSCLC, found the 3-cm cutoff not to be an appropriate prognostic threshold. Four other tumor sizes were better and stronger prognostic indicators: (a) 0 to 2 cm, (b) 2.1 to 4 cm, (c) 4.1 to 7 cm, and (d) >7 cm. Obviously more data are necessary to resolve this issue.

As to the two major cell types of small peripheral lung tumors, squamous cell carcinomas comprise only approximately 5% of such tumors, according to the study of Sakurai and colleagues.^243,244 Of interest is that the number of peripheral lesions represented 22% of all squamous cell carcinomas resected by this group. Twenty-two squamous cell tumors ≤20 mm in size had a 13.6% incidence of lymph node metastasis, which is greater than the incidences of 6.3%, 7.4%, and 0% reported by Asamura,¹² Oda,¹⁹⁰ and Watanabe²⁸⁶ and their colleagues, respectively. In 33 patients with squamous cell tumors 20 to 30 mm in size reported by Sakurai and coworkers,²⁴⁴ the incidence of lymph node metastases was 33%. No skip metatases were seen in either group. The single tumor in the series <10 mm in size had no evidence of metastatic lymph node involvement.