Peripheral Lung Adenocarcinoma: Correlation of Thin-Section CT Findings with Histologic Prognostic Factors and Survival 1

Transcription

1 Thoracic Imaging Takatoshi Aoki, MD Yoshinori Tomoda, MD Hideyuki Watanabe, MD Hajime Nakata, MD Takahiko Kasai, MD Hiroshi Hashimoto, MD Mantaro Kodate, MD Toshihiro Osaki, MD Kosei Yasumoto, MD Peripheral Lung Adenocarcinoma: Correlation of Thin-Section CT Findings with Histologic Prognostic Factors and Survival 1 Index terms: Lung neoplasms, CT, Lung neoplasms, diagnosis, , Lung neoplasms, staging, , , , Published online: August 21, /radiol Radiology 2001; 220: Abbreviations: BAC bronchioloalveolar carcinoma GGO ground-glass opacity LNM lymph node metastasis VI vessel invasion 1 From the Department of Radiology (T.A., Y.T., H.W., H.N.), Department of Pathology and Oncology (T.K., H.H.), and Second Department of Surgery (M.K., T.O., K.Y.), University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu-shi , Japan. Received October 24, 2000; revision requested December 6; final revision received March 27, 2001; accepted April 10. Address correspondence to T.A. ( a-taka@med.uoeh-u.ac.jp) RSNA, 2001 PURPOSE: To evaluate the prognostic importance of thin-section computed tomographic (CT) findings of peripheral lung adenocarcinomas. MATERIALS AND METHODS: The subjects were 127 patients with adenocarcinomas smaller than 3 cm in largest diameter who underwent at least a lobectomy with hilar and mediastinal lymphadenectomy. The margin characteristics of nodules and the extent of ground-glass opacity (GGO) within the nodules at preoperative thin-section CT were analyzed retrospectively. Regional lymph node metastasis (LNM) and vessel invasion (VI) were histologically examined in surgical specimens. Survival curves were calculated according to the Kaplan-Meier method. RESULTS: The frequencies of LNM (4% [1 of 24]) and VI (13% [three of 24]) in adenocarcinomas with GGO components of more than 50% were significantly lower than those with GGO components of less than 10% (LNM, P.05; VI, P.01). The patients with GGO components of more than 50% showed a significantly better prognosis than those with GGO components less than 50% (P.05). All 17 adenocarcinomas smaller than 2 cm with GGO components of more than 50% were free of LNM and VI, and all these patients are alive without recurrence. Coarse spiculation and thickening of bronchovascular bundles around the tumors were observed more frequently in tumors with LNM or VI than in those without LNM or VI (P.01). CONCLUSION: Thin-section CT findings of peripheral lung adenocarcinomas correlate well with histologic prognostic factors. Author contributions: Guarantor of integrity of entire study, T.A., H.N.; study concepts, T.A., H.N.; study design, T.A.; literature research, T.A., Y.T.; clinical studies, Y.T., T.A., H.W., T.K.; data acquisition, T.A., T.K., M.K.; data analysis/interpretation, T.A., Y.T.; statistical analysis, T.A.; manuscript preparation, T.A., H.N.; manuscript definition of intellectual content, H.N., H.H., K.Y., T.O.; manuscript editing, T.A., H.N.; manuscript revision/review, all authors; manuscript final version approval, T.A., H.N. Adenocarcinoma is currently the most predominant histologic type of lung cancer in many countries (1 3). Most patients with early-stage adenocarcinoma are asymptomatic, and their common initial finding is a peripheral nodule detected at radiologic examination (4). In several recent studies (5,6), it was reported that screening with low-dose computed tomography (CT) could improve the detection of lung cancer, especially adenocarcinoma, at an earlier and potentially more curable stage. Therefore, the management of small lung adenocarcinomas has become an important issue. The use of video-assisted thoracic surgery (7) and limited resection for small lung cancer tumors with a favorable prognosis are under examination (8 12), but there is not sufficient evidence to support such an approach. The authors of many studies (13 17) have documented thin-section CT findings of peripheral lung adenocarcinomas. An adenocarcinoma that appears as a localized groundglass opacity (GGO) without spiculation is likely to be a bronchioloalveolar carcinoma (BAC) with slow growth (18). However, the correlation of these thin-section CT findings with survival and known histologic prognostic factors, such as regional lymph node metastasis (LNM) and vessel invasion (VI) (19,20), have not been fully clarified, to our 803

2 Figure 1. Group 3 adenocarcinoma. Transverse thin-section CT scan of lung adenocarcinoma shows peripheral GGO and a central area of increased attenuation. The greatest diameter of the tumor (white line) and the area without GGO (black line) were determined at thin-section CT. The GGO component was calculated as 65%. knowledge. The purpose of this study was to evaluate the prognostic importance of thin-section CT findings of peripheral lung adenocarcinomas. MATERIALS AND METHODS We retrospectively reviewed the records and CT images of 127 patients (68 men, 59 women; age range, years; mean age, 64.0 years) with adenocarcinomas smaller than 3 cm in largest diameter who underwent at least a lobectomy with hilar and mediastinal lymphadenectomy between 1990 and Patients who underwent only resection of enlarged nodes and not mediastinohilar lymphadenectomy at the time of lobectomy were excluded from this study. No patients had evidence of distant metastases at preoperative examinations, including abdominal and brain CT. CT was performed with a helical scanner (TCT-900S, Toshiba Medical Systems, Tokyo, Japan; Somatom Plus, Siemens Medical Systems, Erlangen, Germany). Routine scanning of the entire lung (120 kvp, 150 ma) was first performed in the helical mode with a table speed of 10 mm/sec and 10-mm section thickness. Additional thin-section CT with 2.0-mm section thickness (120 kvp, 250 ma, 1.0second scanning time) to image the tumor was performed for all patients. Thinsection CT images were reconstructed with a high-spatial-frequency algorithm and were printed with fixed window settings (lung center, 700 HU and width, 804 Radiology September 2001 Figure 2. Group 1 adenocarcinoma. (a) Transverse thin-section CT scan shows an mm nodule with solid attenuation without GGO. (b) Histologic specimen reveals solid tumor growth and a distinct boundary (arrows) between the tumor and nontumorous parenchyma. (Hematoxylin-eosin stain; original magnification, 3.) 1,500 HU). The time between CT and surgery ranged from 2 to 48 days. Between September and November 1999, two chest radiologists (H.W., Y.T.) retrospectively analyzed the findings on thin-section CT scans by consensus. They analyzed the margin characteristics of nodules and the internal characteristics within the nodules at preoperative thinsection CT. The tumor contents were semiquantitatively classified according to the extent occupied by GGO within the whole tumor. GGO was defined as hazy and amorphous increased lung attenuation without obscuration of the underlying vascular markings and bronchial walls. The percentage of GGO component was calculated as follows (Fig 1): ([DGGO D]/DGGO) 100, where DGGO is the greatest diameter of the tumor including the GGO area and D is the greatest diameter of the tumor without the GGO area. The tumors were classified into three groups according to the size of the GGO component: less than 10%, group 1; 10% 50%, group 2; and more than 50%, group 3. The observers also recorded the presence of no, fine, or coarse spiculation. According to Zwirewich et al (16), coarse spiculation was defined as the presence of linear strands at least 2 mm thick that extended from the nodule margin into the lung parenchyma, and fine spiculation was defined as the presence of linear strands thinner than 2 mm. Thickening of bronchovascular bundles was considered to have occurred if the normally tapering bundle leading to the nodule was observed to be distinctly widened. CT observers analyzed only the findings on CT scans obtained with lung window settings; they were blinded to the findings on CT scans obtained with mediastinal window settings, pathologic results, and survival data. All surgical specimens were fixed in the inflated state by means of transpleural and transbronchial infusion of formalin. The specimens were stained with hematoxylin and eosin. Regional LNM and VI were independently reviewed in the surgical specimens by a lung pathologist (T.K.) who reviewed all lung specimens again for this study. The diagnosis of regional LNM was based on the TNM classification according to the criteria of the International System for Staging Lung Cancer (21). VI was considered to be present if there was tumor invasion of arteries, veins, or lymphatic vessels. On the basis of the histologic growth pattern, the adenocarcinomas were classified into the following three subtypes: BAC, adenocarcinoma with BAC component, and adenocarcinoma without BAC component. The internal characteristics of the tumors seen on thin-section CT scans were compared with those seen at pathologic examination of the specimens. The correlations were decided by consensus between one pathologist (T.K.) and one radiologist (T.A.). Statistical analysis of the relationship between thin-section CT findings and pathologic findings of LNM or VI was performed by using the Fisher exact test. Survival was calculated from the date of the surgery to the date of death or last contact with the patient. Survival outaoki et al

3 TABLE 1 Relationship between Findings at Thin-Section CT and Pathologic Prognostic Factors GGO Component at CT Thickening of Bronchovascular Bundle at CT Spiculation at CT Pathologic Prognostic Factor 10% (group 1) 10% 50% (group 2) 50% (group 3) Coarse Fine Negative Positive Negative LNM VI 26% (19/73)* 53% (39/73) 17% (5/30) 33% (10/30) 4% (1/24)* 13% (3/24) 53% (8/15)* 73% (11/15)* 17% (8/46)* 41% (19/46)* 14% (9/66) 33% (22/66) 50% (10/20) 85% (17/20) 14% (15/107) 33% (35/107) * P.05, within each category. P.01, within each category. of September 2000 (group 1, four patients; group 2, three patients). Survival curves were calculated according to the Kaplan-Meier method, and statistical evaluation of the factors was performed with the log-rank test by using software (SAS; SAS Institute, Cary, NC). These analyses were performed for all cases with tumors at least 2 cm in largest diameter and also for cases with tumors smaller than 2 cm. We also performed multivariate analyses by using the stepwise Cox proportional hazards model to assess the effects on survival of thin-section CT findings, histologic prognostic factors, sex, and age. Figure 3. Group 2 adenocarcinoma. (a) Transverse thin-section CT scan shows a mm nodule with GGO (arrows) only at the periphery. (b) Histologic specimen shows a replacement growth pattern (arrows) at the periphery of a cellular growth pattern with collagenization (arrowheads). (Hematoxylin-eosin stain; original magnification, 3.) Figure 4. Group 3 adenocarcinoma. (a) Transverse thin-section CT scan shows a mm area completely occupied by GGO (arrows). (b) Histologic specimen reveals a localized BAC growing by means of replacement of alveolar lining cells (arrows). Note the minimal thickening of the alveolar septa and preservation of the lung architecture. (Hematoxylin-eosin stain; original magnification, 10.) comes were obtained from the patient s medical records or from those of their primary care physicians. After surgery, 19 Volume 220 Number 3 (15%) of the 127 patients were treated with chemotherapy or radiation therapy, and seven of the 19 patients were alive as RESULTS The thin-section CT appearances that correlated with LNM and VI are summarized in Table 1. Regarding the GGO component, there were 73 patients in group 1, 30 in group 2, and 24 in group 3 (Figs 2 4). There were no statistically significant differences in patient age and sex among groups on the basis of findings at thin-section CT ( 2 test and unpaired t test, respectively). The frequencies of LNM (4% [1 of 24]) and VI (13% [3 of 24]) in adenocarcinomas in group 3 were significantly lower than those in group 1 (LNM, P.05; VI, P.01). Pathologically, GGO was associated with a growth pattern involving replacement of alveolar lining cells with a relative lack of acinar filling. Coarse spiculation (Fig 5) and thickening of bronchovascular bundles around tumors (Fig 6) occurred with a significantly higher frequency among adenocarcinomas with LNM or VI than among those without LNM or VI (P.05). The desmoplastic response resulted in fibrotic strands, which show direct infiltration of the tumor into adjacent bronchovascular sheaths, and lymphangitic extension corresponded to the coarse spiculation and thickening of the bronchovascular bundles around the tumors. Peripheral Lung Adenocarcinomas at Thin-Section CT 805

4 Direct infiltration of the tumor into adjacent bronchovascular sheaths and lymphangitic extension were identified in seven (47%) of 15 adenocarcinomas with coarse spiculation. The relationship between extent of GGO and tumor subtypes is summarized in Table 2. We used the definition of BAC in the new histologic classification of the 1999 World Health Organization and International Association for the Study of Lung Cancer. If an invasive component was identified in the tumor with bronchioloalveolar growth pattern, it was classified as adenocarcinoma with BAC component. Group 3 consisted of 18 (67%) of 27 pure BAC tumors and six (9%) of 69 adenocarcinomas with BAC component. Postoperative survival rates of the patients with tumors are shown in Figures 7 9. The median follow-up period for the surviving 99 patients was 1,112 days (range, 192 3,627 days). The difference in survival among the three groups according to the GGO component was significant (P.05). Patients in group 3 had a significantly better prognosis than those in groups 1 and 2 (group 3 vs groups 1 and 2, P.05). The presence of coarse spiculation and thickening of the bronchovascular bundles was associated with a significantly lower survival rate (coarse spiculation, P.01; thickening of the bronchovascular bundles, P.01). The relationship between the GGO component and LNM and VI among tumors smaller than 2 cm was also analyzed (Table 3). All 17 patients with adenocarcinomas smaller than 2 cm in group 3 were free of LNM and VI. There was a significant correlation between the three groups according to GGO and VI (P.01). The differences in the frequencies of LNM and VI between adenocarcinomas smaller than 2 cm in group 1 and those in group 3 were significant (LNM, P.05; VI, P.01). All 17 patients with group 3 tumors smaller than 2 cm were alive without recurrence as of September Findings with the Cox proportional hazards model showed that LNM (P.01; odds ratio, 3.26; 95% CI: 1.505, 7.068), coarse spiculation (P.01; odds ratio, 3.17; 95% CI: 1.389, 7.249), and extent of GGO (P.05; odds ratio, 0.063; 95% CI: 0.004, 0.978) were significantly related to survival. Figure 5. Coarse spiculation. (a) Transverse thin-section CT scan depicts coarse spiculation (arrow) around the periphery of the nodule into the lung parenchyma. (b) Histologic specimen shows coarse spiculation is accounted for by direct tumor extension (arrows) along the bronchovascular bundles adjacent to the periphery of the nodule. There was no desmoplastic reaction in this case. (Hematoxylin-eosin stain; original magnification, 10.) Figure 6. Thickening of the bronchovascular bundle. (a) Transverse thin-section CT scan depicts thickening of the bronchovascular bundle (arrow) around the nodule. (b) Histologic specimen shows marked lymphangitic spread of the tumor (arrows) along the bronchus, which corresponds to thickening of the bronchovascular bundle. (Hematoxylin-eosin stain; original magnification, 20.) TABLE 2 Relationship between GGO Component and Tumor Subtype GGO Component Tumor Subtype BAC Adenocarcinoma With BAC component Without BAC component 10% (group 1) 10% 50% (group 2) 50% (group 3) Total Note. Data are the number of lung adenocarcinomas (n 127). DISCUSSION Tumor size in lung cancer is not as good a prognostic factor for adenocarcinoma as it is for squamous cell carcinoma 806 Radiology September 2001 (22,23). Peripheral lung adenocarcinomas smaller than 3 cm frequently recur, resulting in cancer death (5-year disease- free survival for T1N0 disease is approximately 65%) (24). Noguchi et al (25) proposed a simple histologic subclassification Aoki et al

5 Figure 7. Line graph depicts postoperative overall survival according to GGO component (group 1, n 73; group 2, n 30; group 3, n 24). The difference in survival among the three groups according to GGO component was significant (P.05). Figure 8. Line graph depicts postoperative overall survival according to the presence (, n 15) or absence (, n 112) of coarse spiculation. Coarse spiculation was significantly related to survival (P.01). of adenocarcinomas of the lung based on tumor growth patterns; this showed a good correlation with prognosis. They suggested that patients with adenocarcinomas with a growth pattern involving replacement of alveolar lining cells without active fibroblastic proliferation and with no LNM had an excellent outcome after surgical resection. Most of these adenocarcinomas with a good prognosis showed pure GGO or GGO mixed with a small area of solid attenuation at thinsection CT (18,26). Although a tumor with a larger GGO component is likely to be BAC (17,18, 26,27) and therefore to have low propensity for distant spread, it is not possible to differentiate various subtypes of adenocarcinoma distinctively at CT. According to recent reports (18,27) of analysis of serial changes in the appearance of BAC at CT, localized GGO can change into mixed areas of GGO and solid attenuation, and the solid component can increase during the interval. These results prompted us to suspect that an adenocarcinoma with a larger GGO area is at an early and curable stage and that thinsection CT findings may provide a clue to predict the prognostic outcome for patients with lung adenocarcinomas. Several authors (16,17) have reported that the GGO at thin-section CT in pathologically proved lung adenocarcinomas corresponds to a growth pattern involving alveolar septa with a relative lack of acinar filling. Our results showed the GGO component was well correlated with histologic prognostic factors, such as LNM, VI, and survival. The frequencies of LNM and VI in adenocarcinomas with a GGO component of more than 50% were significantly lower than those with a GGO component of less than 10%. Although distinction between pure BAC and adenocarcinoma with BAC component could not be made at thin-section CT, the patients with a GGO component of more than 50% had a significantly better prognosis. When tumors with GGO components of more than 50% were limited to a diameter smaller than 2 cm, no LNM or VI was seen, and the outcome was excellent. Therefore, the GGO component of tumors can be an excellent predictor of the histologic prognostic factors and survival in small peripheral lung adenocarcinoma, and it may allow differentiation of adenocarcinomas with various biologic behaviors. Zwirewich et al (16) showed that coarse spiculation occurred with a significantly higher frequency among malignant lung nodules than among benign lung nodules. They reported that the spiculation pathologically correlated with a desmoplastic response in the nodule, which resulted in fibrotic strands radiating into the surrounding lung parenchyma, or with direct infiltration of the tumor into adjacent bronchovascular sheaths or localized lymphangitic extension. These findings were not distinguishable on thin-section CT images. In this study, coarse spiculation and thickening of bronchovascular bundles around tumors occurred with a significantly higher frequency among adenocarcinomas with LNM or VI than among those without Volume 220 Number 3 Peripheral Lung Adenocarcinomas at Thin-Section CT 807

6 LNM or VI, and they were significantly related to survival. Although several conditions resulting in chronic interstitial fibrosis of the lungs are associated with the later development of cancer, most of the so-called scars in adenocarcinomas are currently considered a desmoplastic reaction to the tumor and are formed during tumor growth (22,28). Shimosato et al (22) demonstrated that central fibrosis occurs as a result of alveolar collapse secondary to degeneration and death of tumor cells lining the alveoli and that a high degree of collagenization is correlated with a poor outcome. Although not all the coarse spiculation corresponded to direct infiltration of the tumor into adjacent bronchovascular sheaths or localized lymphangitic extension, the desmoplastic response resulting in coarse fibrotic strands radiating from the nodule may suggest an advanced adenocarcinoma with a high degree of collagenization of surrounding lung parenchyma. Since small adenocarcinomas of the lung have been increasingly detected in elderly patients with recent advances in diagnostic modalities, a suitable surgical approach that achieves the most benefit for these patients must be considered. Several studies (11,29) dealt with the risk of limited surgery in patients with lung cancer. However, limited surgical resection has the benefit of preserving the postoperative quality of life without impairment of respiratory function (30). Although such a surgical approach for lung cancer tumors smaller than 3 cm has not shown the same effectiveness as standard surgery, including lobectomy (11), limited surgical resection in properly selected cases could offer the same therapeutic effect. CT calculation software could not be used in this retrospective study, because the GGO component had been calculated by using a semiquantitative method. The results of our study show that a lung adenocarcinoma smaller than 2 cm with a GGO component of more than 50% at thin-section CT has a high likelihood of being free of LNM or VI. In clinical settings in which limited surgical resection is desirable, preoperative quantification of GGO at thin-section CT may be a helpful tool in determining eligibility of patients. Acknowledgments: The authors thank Shinya Matsuda, MD, and Tsutomu Housyuyama, MD, for their advice regarding statistical analysis. Figure 9. Line graph depicts postoperative overall survival according to the presence (, n 20) or absence (, n 107) of thickening of the bronchovascular bundle. Thickening of the bronchovascular bundle was significantly related to survival (P.01). TABLE 3 Relationship between GGO Component of Lung Adenocarcinoma Smaller than 2 cm and Pathologic Prognostic Factors Pathologic Prognostic Factor 10% (group 1) References 1. Barsky SH, Cameron R, Osann KE, Tomita D, Holmes EC. Rising incidence of bronchioloalveolar lung carcinoma and its unique clinicopathologic features. Cancer 1994; 73: Travis WD, Travis LB, Devesa SS. Lung cancer. Cancer 1995; 75: Auerbach O, Garfinkel L. The changing pattern of lung carcinoma. Cancer 1991; 68: Edgerton F, Rao U, Takita H, Vincent RG. Bronchio-alveolar carcinoma: a clinical overview and bibliography. Oncology 1981; 38: Sone S, Takashima S, Li F, et al. Mass screening for lung cancer with mobile spiral computed tomography scanner. Lancet 1998; 351: Henschke CI, McCauley DI, Yankelevitz DF, et al. Early lung cancer action project: overall design and findings from baseline screening. Lancet 1999; 354: Landreneau RJ, Mack MJ, Hazelrigg SR, et al. Video-assisted thoracic surgery: basic technical concepts and intercostal approach strategies. Ann Thorac Surg 1992; 54: Miller JI, Hatcher CR Jr. Limited resection GGO Component 10 50% (group 2) 50% (group 3) Total LNM 23% (10/44)* 17% (2/12) 0% (0/17)* 16% (12/73) VI 54% (23/44) 42% (5/12)* 0% (0/17)* 38% (28/73) * P.05, within each category. P.01, within each category. of bronchogenic carcinoma in the patient with marked impairment of pulmonary function. Ann Thorac Surg 1987; 44: Read RC, Yoder G, Schaeffer RC. Survival after conservative resection for T1 N0 M0 non-small cell lung cancer. Ann Thorac Surg 1990; 49: Pastorino U, Valente M, Bedini V, Infante M, Tavecchio L, Ravasi G. Limited resection for stage I lung cancer. Eur J Surg Oncol 1991; 17: Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1N0 non-small cell lung cancer. Ann Thorac Surg 1995; 60: Yoshida J, Nagai K, Yokose T, et al. Primary peripheral lung carcinoma smaller than 1 cm in diameter. Chest 1998; 114: Webb WR. Radiologic evaluation of the solitary pulmonary nodule. AJR Am J Roentgenol 1990; 154: Adler B, Padley S, Miller RR, Muler NL. High-resolution CT of bronchioloalveolar carcinoma. AJR Am J Roentgenol 1992; 159: Siegelman SS, Khouri NF, Leo FP, Fishman EK, Braverman RM, Zerhouni EA. 808 Radiology September 2001 Aoki et al

7 Solitary pulmonary nodules: CT assessment. Radiology 1986; 160: Zwirewich CV, Vedal S, Miller RR, Muler NL. Solitary pulmonary nodules: highresolution CT and radiologic-pathologic correlation. Radiology 1991; 179: Kuhlman JE, Fishman EK, Kuhajda FP, et al. Solitary bronchioloalveolar carcinoma: CT criteria. Radiology 1988; 167: Aoki T, Nakata H, Watanabe H, et al. Evolution of peripheral lung adenocarcinomas: CT findings correlated with histology and tumor doubling time. AJR Am J Roentgenol 2000; 174: Fujisawa T, Yamaguchi Y, Saito Y, Hiroshima K, Ohwada H. Blood and lymphatic vessel invasion as prognostic factors for patients with primary resected nonsmall cell carcinoma of the lung with intrapulmonary metastases. Cancer 1995; 76: Macchiarini P, Fontanini G, Hardin MJ, et al. Blood vessel invasion by tumor cells predicts recurrence in completely resected T1N0M0 non-small-cell cancer. J Thorac Cardiovasc Surg 1993; 106: Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997; 111: Shimosato Y, Suzuki A, Hashimoto T, et al. Prognostic implication of fibrotic focus (scar) in small peripheral lung cancer. Am J Surg Pathol 1980; 4: Gail MH, Eagan RT, Feld R, et al. Prognostic factors in patients with resected stage I non-small cell lung cancer: a report from the lung cancer study group. Cancer 1984; 54: Thomas PA, Piantadosi S. Postoperative T1N0 non-small cell lung cancer: squamous versus nonsquamous recurrences. J Thorac Cardiovasc Surg 1987; 94: Noguchi M, Morikawa A, Kawasaki M, et al. Small adenocarcinoma of the lung. Cancer 1995; 75: Kuriyama K, Seto M, Kasugai T, et al. Ground-glass opacity on thin-section CT: value in differentiating subtypes of adenocarcinoma of the lung. AJR Am J Roentgenol 1999; 173: Jang HJ, Lee KS, Kwon OJ, Rhee CH, Shim YM, Han J. Bronchioloalveolar carcinoma: focal area of ground-glass attenuation at thin-section CT as an early sign. Radiology 1996; 199: Barsky SH, Huang SJ, Bhuta S. The extracellular matrix of pulmonary scar carcinomas is suggestive of a desmoplastic origin. Am J Pathol 1986; 124: Ichinose Y, Yano T, Yokoyama H, et al. The correlation between tumor size and lymphatic vessel invasion in resected peripheral stage I non-small cell lung cancer. J Thorac Cardiovasc Surg 1994; 108: Stair JM, Womble J, Schaefer RF, Read RC. Segmental pulmonary resection for cancer. Am J Surg 1985; 150: Volume 220 Number 3 Peripheral Lung Adenocarcinomas at Thin-Section CT 809