ORIGINAL ARTICLES
Chest Wall Involvement by Lung Cancer: Computed Tomographic Detection and Results of Operation G. B. Ratto, MD, G. Piacenza, MD, C. Frola, MD, F. Musante, MD, I. Serrano, MD, R. Giua, MD, M. Salio, MD, P. Jacovoni, MD, and S. Rovida, MD Institute of Clinica Chimrgica, University of Genoa, Genoa, Italy
The aim of this prospective study was to evaluate: (1)the role of computed tomographic scanning in predicting chest wall invasion by peripheral lung cancer and (2) the results of operation according to the depth of chest wall involvement and other potential indicators of long-term survival. One hundred twelve patients with non-small cell lung cancer adjacent to the pleural surface who underwent computed tomographic scanning and subsequent thoracotomy were entered into this study. Tumor invasion was confined to the visceral pleura in 53 patients, to the parietal pleura in 18 patients, and to intercostal muscles in 25 patients; invasion extended beyond this layer in 16 patients. The computed tomographic criteria for chest wall invasion were (1)obliteration of the extrapleural fat plane, (2) the length of the
tumor-pleura contact, (3) the ratio between the tumorpleura contact and the tumor diameter, (4) the angle of the tumor with the pleura, (5) a mass involving the chest wall, and (6)rib destruction. The computed tomographic criteria 1 and 3 were significantly related to pathologic findings. Sensitivity was 85% for criterion 1 and 83% for criterion 3, specificity being 87% and 80%, respectively. Long-term survival of patients with T3 disease critically depended on the lymph node state and completeness of resection. The adenocarcinoma cell type and the T4 category were unfavorable prognostic factors. The depth of chest wall invasion did not affect survival, except for extensive rib and soft tissue infiltration. En bloc resection yielded better results than discontinuous resection. (Ann Thoruc Surg 2992;52:282-8)
S
Patients and Methods
urgeons have now recognized that patients with nonsmall cell lung cancer invading the chest wall have a potential long-term survival, provided that complete resection is feasible and mediastinal node involvement is not present [l-31. The decision whether to undertake en bloc resection of the lung and chest wall or to attempt extrapleural dissection depends on the presumed degree of peripheral tumor spread. Even intraoperatively, differentiation between neoplasms confined within the parietal pleura from those infiltrating the chest wall proper may be difficult. Preoperative assessment of chest wall involvement is therefore important in planning treatment. Although computed tomographic (CT) scanning has been suggested to be less than adequate to establish the depth of chest wall infiltration (4-61, it remains the most accurate diagnostic procedure currently available in Italy. The present prospective study was carried out to evaluate (1) the diagnostic accuracy of individual CT criteria and their combinations in predicting the extent of chest wall invasion and (2) the results of operation according to the depth of chest wall involvement and other potential indicators of long-term success.
Accepted for publication Sep 26, 1990. Address reprint requests to Dr Ratto, Istituto di Clinica Chirurgica la University of Genoa, Viale Benedetto XV, 10, 16132 Genoa, Italy.
0 1991 by The
Society of Thoracic Surgeons
From January 1983 through December 1988, patients admitted to four institutions (Clinica Chirurgica Ia, University of Genoa; San Martino and Sanpierdarena Hospitals, Genoa; and Alessandria Hospital, Alessandria) with peripheral, potentially resectable lung cancer were entered into this prospective study. The inclusion criteria were (1) histologic diagnosis of non-small cell lung cancer, (2) preoperative CT scanning demonstrating contiguity of the tumor with the pleural surface, and (3) thoracotomy performed. Subjects who had superior sulcus neoplasms, malignant pleural effusions, or preoperative radiotherapy and those who were not available to follow-up were excluded from the study. Computed tomographic images were read independently by two of us (C.F. and F.M.) before operation. They were asked if the tumor was confined within the parietal pleura or invaded extrapleural soft tissues. Preoperative staging procedures included chest roentgenography, bronchoscopy, and computed tomography of the chest and abdomen. Bone and brain scans, as well as cervical mediastinoscopy, were not included in the staging protocol. Throughout the entire study duration, the follow-up of patients was carried out according to Ginsberg’s suggestions [7]. There were 102 men and 10 women. The mean age was 8.6 years. Histology of the primary tumor was 59.8 squamous cell carcinoma (47 patients), adenocarcinoma (56), large cell carcinoma (5), and adenosquamous carci-
*
0003-4975/91/$3.50
Ann Thorac Surg 1991;Sl:182-8
noma (4). At operation, 13 neoplasms proved to be unresectable because of invasion of the spine, unremovable node metastases, or aortic arch infiltration. In 13 patients the resection was classified as incomplete because microscopic examination showed either the margins of the resected specimen or the most distally removed node in major lymphatic drainage regions to be infiltrated by tumor cells. Of the 86 subjects who underwent a complete resection, 21 had pneumonectomy and 65 lobectomy or bilobectomy. Biopsies of attainable paratracheal, subcarinal, inferior pulmonary ligament, and hilar nodes were routinely performed. The same surgical policy was applied by the surgeons participating in the study (G.B.R., I.S., R.G., and P.J.). The decision to perform an extrapleural dissection or en bloc resection depended on the operative findings. A trial of extrapleural dissection was attempted when there was no clear evidence of tumor extension into the intercostal muscles or ribs. This approach was used in 76 patients: 18 had parietal pleura invasion and 19 had tumor extension beyond the parietal pleura. In these latter patients, the involved chest wall was discontinuously resected. If during extrapleural dissection there was strong fixation of the parietal pleura to the overlying ribs, en bloc resection was carried out (18 patients). The number of resected ribs ranged from one to four. Prosthetic materials were never used for parietal reconstruction. Nineteen neoplasms were postsurgically classified as stage I, 9 as stage 11, 66 as stage IIIa, and 18 as stage IIIb. Histologic examination of the removed specimens demonstrated tumor involvement confined to the visceral pleura in 53 patients, parietal pleura involvement in 18, and intercostal muscle or rib periostium involvement in 25. In 16 patients there was either soft tissue infiltration beyond the intercostal muscles or rib destruction. Postoperative radiation therapy was given to patients with known residual tumor. Two patients died within 30 days after operation because of respiratory failure and intrathoracic bleeding. No death followed en bloc or discontinuous chest wall resection.
Computed Tomographic lnvestigations Computed tomographic scans of the chest were obtained using a Somatom I1 (Siemens, Erlangen, Germany) scanner. Contiguous cuts, 8 mm thick, were taken at full inspiration from the diaphragm to the lung apexes without overlap. Contrast medium was infused intravenously. The CT scanning criteria used to predict the depth of chest wall involvement were (1)obliteration of the extrapleural fat plane, (2) the length of the tumor contact with the adjacent pleural surface, (3) the ratio between the amount of tumor-pleura contact and the tumor diameter, (4) the angle of the tumor with the pleura (acute or obtuse angle), (5) an obvious mass extending into the chest wall soft tissues, and (6) rib destruction. Thickening of the pleura adjacent to the neoplasm was not taken into account because it has been previously demonstrated to be nonspecific [4]. The extrapleural fat plane was not uniformly discernable, particularly on CT cuts corresponding to the rib level. Nevertheless, the plane was practically always
RA'ITOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
183
Fig I. Right bronchogenic carcinoma contiguous with the chest wall. The extrapleural fat plane is well visible, with the arrow showing encroachment of tumor on extrapleural fat. Irregularity of the rib border, suggesting bone erosion, is also demonstrated.
visible if researched between the pleura adjacent to the tumor (which was often thickened) and the intercostal muscle layer (Fig 1).Measurement of tumor diameterwas obtained from CT scans using the soft-tissue window settings [4].Differentiation between cancer and surrounding atelectasis or pneumonitis was attempted by ahalyzing the vascularization patterns after contrast medium injection. The contact between abnormal lung and chest wall was assessed as shown in Figure 2. Rib destruction was searched for using the thin-cut technique and was defined as any evidence of irregular rib contour (see Fig 1). When the angle of the tumor with the pleura was either acute or obtuse in different areas, it was classified as obtuse.
Statistical Analysis Computed tomographic findings were correlated with pathologic findings, which were assumed as the true reference. Comparison was made in terms of sensitivity, specificity, and accuracy according to standard formulas. The predictive value of CT results was assessed both for each CT criterion and for their combinations. When a combination of several CT criteria was used, the diagnostic efficacy of the test was evaluated in two ways: (1) by classifying the test positive if at least one CT criterion indicated tumor infiltration beyond the parietal pleura, and (2) by considering the test positive if all the selected CT criteria suggested chest wall invasion. The strength of the association between the results provided by CT scanning and by thoracotomy was evaluated by multiple regression analysis. Because CT criteria 1, 5, and 6 were
184
RATTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Ann Thorac Surg 1991;51:182-8
Table 2. Diagnostic Efficacy of Computed Tomographic Criterion 2 (Length of Tumor Contact With Pleural Surface) According to Different Cut-off Values for Separating Patients With Chest Wall Invasion From Those Without Tumor-Pleura Contact (cm) Value Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (%)
Fig 2. L-$t bronchogenic carcinoma infiltrating the chest wall. The technique used for measuring the length of tumor in contact with pleura is illustrated. subjective in nature, intraobserver and interobserver variation have been estimated by means of the 4 coefficient method. This coefficient was 0.03 for intraobserver variation and 0.05 for interobserver variation. Survival rates have been calculated by the Kaplan-Meier method. Comparison of survival curves was done by the MantelHaensel test.
Results Sensitivity, specificity, and accuracy of individual CT criteria for the diagnosis of chest wall invasion, as well as the predictive value of positive or negative tests, are reported in Table 1. Sensitivity, specificity, and accuracy rates of CT criterion 2 (the length of the tumor contact with the adjacent pleural surface) when using different cut-off values separating patients with chest wall infiltration from those without infiltration are listed in Table 2. Sensitivity, specificity, and accuracy rates of CT criterion 3 Table 1. Computed Tomographic Scan Criteria for Chest Wall Invasion
Value Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (%)
Obliteration of the Chest Extrapleural Obtuse Wall Rib Fat Plane Angle Mass Destruction 85 87 86 80
57 72 66 55
33 100 75 100
16 98 67 87
91
73
71
66
97 42 61 51
92 54 67 54
76 68 71 59
97
93
83
(the ratio between the length of tumor-pleura contact and the tumor diameter) when adopting various cut-off values are summarized in Table 3. The correlation between CT data and pathologic findings was significant for CT criteria 1 ( p < 0.001) and 3 ( p < 0.01). Sensitivity, specificity, and accuracy rates of combining CT criteria 1 and 3 are shown in Table 4. The patients’ age and sex, the tumor diameter, and the number of resected ribs did not significantly affect survival. Survival rates depended on the tumor stage ( p < 0.05) and histology ( p < 0.05), the length of tumor-pleura contact ( p < 0.05), and the completeness of resection ( p < 0.001). The 5-year survival rate, according to the disease stage, was 48% in stage I, 31% in stage 11, 12% in stage IIIa, and 0% in stage IIIb. Survival curves of patients with T3 tumors, according to the N status, are illustrated in Figure 3. Patients with cancers classified as T3 NO survived longer than those with cancers classified as T3 N2 ( p < 0.05); patients with T3 N1 lesions also survived longer than those with T3 N2 lesions ( p < 0.05). Similar conclusions were obtained when the analysis was restricted to the individuals who had undergone a complete resection. The 5-year survival rate, according to the tumor cell type, was 25% in subjects with squamous cell carcinoma and Table 3 . Diagnostic Efficacy of Computed Tomographic Criterion 3 (Ratio Between Tumor-Pleura Contact and Tumor Diameter) According to Different Cut-off Values for Separating Patients With Chest Wall Invasion From Those Without Tumor-Pleura ContactTumor Diameter Value
0.5
0.7
0.9
Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (a)
100
37 60 48
92 62 74 60
83 80 81 71
100
94
89
1991;51:182-8
Table 4 . Combined Use of Computed Tomographic Criteria 1 and 3 (Adopting 0.7 as Cut-of Value) for Predicting Chest Wall lnvasion Value
A
B
Sensitivity (%) Specificity (%)
97 61 75 60
81 89 86 81
98
89
Accuracy (%)
Positive predictive value (%) Negative predictive value (%)
185
RA'ITOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Ann Thorac Surg
100
90 80
70 60 50
A = test classified as positive if at least one criterion indicates chest wall invasion; B = test classified as positive if both criteria indicate chest wall invasion.
2o 10
17% in those with adenocarcinoma. When limiting the study to patients who had a potentially complete resection, the relevant figures were 30% and 19%, respectively. The length of the tumor-pleura contact affected survival exclusively when lesions adhering less than 3 cm were compared with lesions adhering more than 6 cm (either if considering the entire series or just the completely resected cases). The 5-year survival rate of patients with T3 tumors who underwent a complete resection (45 patients) was 16%. There were no 2-year survivors in the group of patients with T3 tumors who had incomplete resection or no resection (10 patients). Survival rates of patients with tumors extending to or beyond the parietal pleura who had undergone an extrapleural dissection, en bloc resection, or discontinuous resection of the involved lung and chest wall are shown in Figure 4. There was no significant difference in survival between patients of the extrapleural dissection group and
1 6
12 18 24
30 36 42 48 54 60
MONTHS
Fig 4 . Survival curves of patients with tumors extending to or beyond extrathe parietal pleura, according to the type of resection. (U en bloc resection, 18 patients; pleural dissection, 18 patients; 0-0 discontinuous resection, 19 patients.)
*-*
those of the en bloc resection group. Survival rates after a discontinuous resection were significantly poorer than those after en bloc resection ( p < 0.01). Survival rates according to the depth of chest wall infiltration are illustrated in Figure 5. The probability of long-term survival for patients with cancers confined to the visceral pleura was significantly greater than that of patients with rib destruction or soft tissue infiltration beyond the intercos-
x 100 90 80
70
60 50 40
30 20
\
10
1°1 O
6 12 18 24 30 36 42 48 54 60
HONTHS
Fig 5 . Survival curves of patients with tumors contiguous with the pleural surface, according to the depth of chest wall infiltration. 6 12 18 24 30 36 42 48 54 60 MONTHS (M tumor involvement confined within the visceral pleura, 53 patients; Hparietal pleural involvement, 18 patients; Fig 3 . Survival curves of patients with tumors extending to or beyond the parietal pleura (T3 tumors), according to the N status. (U intercostal muscle or rib periostial involvement, 25 patients; rib destruction or chest wall infiltration beyond intercostal muscles, 16 T3 NO, 14 patients; UT3 N1, 19 patients; T3 N2, 22 patients.) patients.)
0
*---*
*-*
*-*
186
RATTO ET AL CHEST WALL INVOLVEMENT BY LUNG CANCER
tal muscles ( p < 0.01 for the entire series; p < 0.05 for completely resected cases). No significant difference in survival was found among patient groups with intermediate degrees of chest wall infiltration.
Comment Preoperative diagnosis of chest wall invasion by primary lung cancer is important for surgical strategy. Mortality and morbidity associated with unnecessary extended resections [%lo] or unneeded violations of the tumor plane could be avoided. Combined therapy programs, including the use of preoperative radiotherapy, might be prospectively applied. Patients with chest wall invasion and mediastinal node metastases, who do not seem to benefit from en bloc resection [8, 10, 111, could be excluded from radical operations. Although CT scanning has been recognized as a more accurate diagnostic procedure than standard tomography [12-151, it has been recently suggested to be unsuitable to differentiate tumor contiguity with the pleural surface from definite chest wall invasion [ P 6 , 161. The individual criteria used to predict chest wall infiltration proved to be either highly sensitive but nonspecific or highly specific but insensitive. Among the criteria adopted in the present study, obliteration of the extrapleural fat plane was the most significantly related to pathologic findings. Both sensitivity (85%) and specificity (87%)rates were acceptable. The major criticism to the use of this criterion is the variable thickness of the extrapleural fat plane, often indistinguishable on CT images. In our experience, the fat plane between the pleura adjacent to or invaded by the tumor and the intercostal muscle layer was virtually always visible. One could object that, by focusing the analysis to the intercostal muscle regions, rib invasion without obvious bone destruction might have been overlooked. Actually, this occurred in 3 patients only, suggesting that the event is not common. Whether radionuclide bone scanning could be useful in identifying patients with rib involvement undetectable on CT scans remains to be clarified. A false diagnosis of extrapleural fat plane obliteration was made in 9 patients. According to Pennes and associates [4] and Shin and co-workers [17], the relevant causes have been identified in concomitant inflammatory processes of the pleura extending into the extrapleural fat or in chest asymmetry due to the patient’s position. Besides obliteration of the extrapleural fat plane, the only CT criterion significantly related to pathologic data was the ratio between the tumor-pleura contact and tumor diameter. One of the reasons accounting for this finding might reside in the fact that measurement errors were in part withdrawn by expressing the final results as a ratio. If surrounding atelectasis or pneumonitis would be erroneously deemed cancer, inaccuracy might affect the size of both the tumor diameter and tumor-pleura contact. All the remaining CT criteria that have been used in this study were not significantly associated with pathologic data. In an attempt to enhance the predictive value of CT findings, we combined the results of criteria 1 and 3. By using such an association, the following information
Ann Thorac Surg 1991;51:182-8
could be obtained: (1) if both criteria are negative, the surgeon may be almost certain that an extrapleural dissection will not violate the cancer plane (sensitivity: 97%); (2)if both criteria suggest chest wall invasion, the surgeon will be advised against extrapleural dissection (specificity: 89%). The relationship between the depth of chest wall infiltration and prognosis is one of the still-unsolved problems. Gronqvist and associates [18] and Mishina and colleagues [19] suggested that invasion of the pleura is a more favorable indicator than invasion of intercostal muscles, whereas rib involvement gives the worst prognosis. Neef [20] and Martini (211 claimed that patients with invasion of the chest wall proper have a poorer prognosis than patients with invasion of the parietal pleura only. By contrast, McCaughan and co-workers [l]found that the extent of chest wall involvement does not significantly affect survival. The present study showed that tumor infiltration confined within the visceral pleura is significantly associated with a better outcome if compared with infiltration of chest wall soft tissues beyond the intercostal muscles or rib destruction. No significant difference in survival was demonstrated between patients with parietal pleura involvement and those with involvement of the chest wall proper. The observation remained valid even when corrected for the completeness of resection and nodal metastases. However, it should be noted that patients with parietal pleura invasion or chest wall invasion were unevenly treated. The chosen operation in subjects with tumor growth confined to the parietal pleura was extrapleural dissection, whereas subjects with invasion beyond the parietal pleura underwent en bloc or discontinuous chest wall resection. Controversy continues concerning the value of potential indicators of long-term success in patients with lung cancer abutting the chest wall [l,2, 11,221. In accordance with previous reports [l, 10, 11, 191, the present study confirmed that long-term survival is highly dependent on the lymph node state and completeness of operation. The poor results we have obtained in patients with T3 N2 disease further stress the need for mediastinoscopic exploration in individuals whose CT scanning shows chest wall infiltration and enlarged mediastinal nodes [ll]. The lack, in our experience, of 2-year survivors after incomplete resection and radiotherapy does not support the efficacy of combining these therapeutic modalities. The other prognostic factors we found to be correlated with long-term survival were tumor histology, the T component, and the type of resection. Patients with squamous cell carcinoma had a better outcome than those with adenocarcinoma. This finding might suggest that the diagnostic approach to patients ’ with adenocarcinoma invading the chest wall should include a meticulous search for distant metastases to determine if they are actually MO. The erosion of survival expectation according to the T categories was significantly higher for T4 tumors than for T2 or T3 tumors. This would indicate that patients with vertebral body or great vessel invasion are poor candidates for surgical resection [2]. The prognostic value of the type of resection has been
Ann Thorac Surg 1991;51:182-8
differently appreciated by authors. Grillo [ll] found 3 long-term survivors in a group of 19 patients who had undergone en bloc resection, and none in a group of 14 patients in whom extrapleural dissection of a tumor involving the parietal pleura was done. Trastek and associates [23], analyzing subjects with lung cancer infiltrating the parietal pleura, noted 5-year survival rates of 75% after en bloc resection and 27.9% after extrapleural dissection. McCaughan and colleagues [ 11only attempted en bloc resection when there was operative evidence of intercostal muscle or rib invasion, and they concluded that survival is superior in patients undergoing extrapleural dissection (48%) than in those who had en bloc resection (16%). Casillas and co-workers [24] reported no significant difference in survival between the en bloc resection (21% at 5 years) and the extrapleural dissection (24%)groups. In the present series, the survival rates of patients who underwent en bloc resection were not statistically different from those of patients who underwent extrapleural dissection. Nevertheless, there was a tendency toward a better outcome in the first group than in the second group. This in spite of the more extended chest wall invasion characterizing subjects who required en bloc resection. We found survival to be significantly impaired when a discontinuous resection was carried out in place of en bloc resection. This observation is at variance with a previous report [lo] suggesting that discontinuous resection may be adequate in patients with tumor extension beyond the parietal pleura. The explanation might be that, in the earlier series, discontinuous resection was applied only when the adherence of the lung to the chest wall was neither solid nor extensive, whereas this was not always true in the present series. In conclusion, among the numerous CT criteria used for preoperative assessment of chest wall invasion by primary non-small cell lung cancer, obliteration of the extrapleural fat plane and the tumor-pleura contact-tumor diameter ratio were the only variables significantly related to pathologic findings. The depth of chest wall infiltration did not affect long-term survival, provided that extensive invasion of chest wall muscles or ribs was not present. Survival was critically dependent on the lymph node state and completeness of resection. The adenocarcinoma cell type and the T4 category were unfavorable prognostic factors. In patients with tumors extending beyond the parietal pleura, en bloc resection yielded better success rates than discontinuous chest wall resection. Whether en bloc resection is more advantageous than extrapleural dissection in patients with neoplasms confined t o the parietal pleura needs further investigation. The assistance of M. Galanti in the preparation of the manuscript is gratefully acknowledged.
References 1. McCaughan B, Martini N, Bains MS, McCormack PM. Chest
wall invasion in carcinoma of the lung. J Thorac Cardiovasc Surg 1985;89:83641.
RAlTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
187
2. Mountain CF. Value of the new TNM staging system for lung cancer. Chest 1989;96:47S-9S. 3. Miller KB, Little AG. Stage 111 lung cancer. Is surgery appropriate? Pulm Perspect 1989;6(2):2-5. 4. Pennes DR, Glazer GM, Wimbish KJ, Gross BH, Long RW, Orringer MB. Chest wall invasion by lung cancer: limitations of CT evaluation. AJR 1985;144:507-11. 5. Haggar AM, Pearlberg JL, Froelich JW, et al. Chest wall invasion by carcinoma of the lung. Detection by MR imaging. AJR 1987;148:1075-8. 6. Scott IR, Muller NL, Miller RR, Evans KG, Nelems 8. Resectable stage 111 lung cancer: CT, surgical, and pathologic correlation. Radiology 1988;166:75-9. 7. Ginsberg RJ. Follow-up supervision after resection for lung cancer. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery; vol 1: Lung cancer. Philadelphia: W.B. Saunders, 1985:274. 8. Piehler JM, Pairolero PC, Weiland LH, Offord KP, Payne WS, Bernatz PE. Bronchogenic carcinoma with chest wall invasion: factors affecting survival following en bloc resection. Ann Thorac Surg 1982;34:68&91. 9. Paone JF, Spees EK, Newton CG, Lillemoe KD, Kieffer RF, Gadacz TR. An appraisal of en bloc resection of peripheral bronchogenic carcinoma involving the thoracic wall. Chest 1982;81:203-7. 10. Patterson GA, Ilves R, Ginsberg RJ, Cooper JD, Todd TRJ, Pearson FG. The value of adjuvant radiotherapy in pulmonary and chest wall resection for bronchogenic carcinoma. Ann Thorac Surg 1982;34:692-7. 11. Grillo HC. Technical considerations in stage I11 disease: pleural and chest wall involvement. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery; vol 1: Lung cancer. Philadelphia: W.B. Saunders, 1985:134. 12. Ekholm S, Albrechtson U, Kugelberg J, Tylen U. Computed tomography in preoperative staging of bronchogenic carcinoma. J Comput Assist Tomogr 1980;4:76>5. 13. Baron LR, Levitt RG, Sagel SS, White MJ, Roper CL, Marbarger JP. Computed tomography in the preoperative evaluation of bronchogenic carcinoma. Radiology 1982;145: 727-32. 14. Epstein DM, Stephenson LW, Gefter WB, Van der Voorde F, Aronchik JM, Miller JT. Value of CT in the preoperative assessment of lung cancer: a survey of thoracic surgeons. Radiology 1986;161:423-7. 15. Rendina EA, Bognolo DA, Mineo TC, et al. Computed tomography for the evaluation of intrathoracic invasion by lung cancer. J Thorac Cardiovasc Surg 1987;94:57-63. 16. Glazer HS, Duncan-Meyer J, Aronberg DJ, Moran JF, Levitt RG, Sagel SS. Pleural and chest wall invasion in bronchogenic carcinoma: CT evaluation. Radiology 1985;157 1914 17. Shin MS, Anderson SD, Myers J, Ho KJ. Pitfalls in CT evaluation of chest wall invasion by lung cancer. J Comput Assist Tomogr 1986;10136-8. 18. Gronqvist YKJ, Clagett OT, MacDonald JR. Involvement of thoracic wall in bronchogenic carcinoma: study of 16 cases in which pneumonectomy or lobectomy and simultaneous resection of the thoracic wall were done. J Thorac Surg 1957; 33:487-95. 19. Mishina H, Suemasu K, Yoneyama T. Surgical pathology and prognosis of the combined resection of chest wall and lung in lung cancer. Jpn J Clin Oncol 1978;8:161-8. 20. Neef H. Discussion of CasiIlas M, Pans F, Tarrazona V,
188
RAlTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Padilla J, Paniagua M, Galan G. Surgical treatment of lung carcinoma involving the chest wall. Eur J Cardiothorac Surg 1989;3:425-9. 21. Martini N. Discussion of: Patterson GA, Ilves R, Ginsberg RJ, Cooper JD, Todd TRJ,Pearson FG. The value of adjuvant radiotherapy in pulmonary and chest wall resection for bronchogenic carcinoma. Ann Thorac Surg 1982;34:692-7. 22. Sellman M, Heme A, Peterffy A. Extended intrathoracic
Ann Thorac Surg 1991;51:182-8
resection for lung cancer. Scand J Thorac Cardiovasc Surg 1987;21:69-72. 23. Trastek VF, Pairolero PC, Piehler JM, et al. En bloc (non-chest wall) resection for bronchogenic carcinoma with parietal fixation. J Thorac Cardiovasc Surg 1984;87352-8. 24. Casillas M, Paris F, Tarrazona V, Padilla J, Paniagua M, Galan G. Surgical treatment of lung carcinoma involving the chest wall. Eur J Cardiothorac Surg 1989;3:425-9.
IrnDortant Announcement for Authors Elsevier Science Publishing Company has recently instituted electronic publishing for certain of their journals, including The Annals of Thoracic Surgery. Beginning September 1, 1990, the final version of accepted manuscripts submitted on diskettes will be converted electronically directly to page proof, with a considerable saving of time a n d expense. Authors are encouraged to use this option if they wish to do so. Both 3.5"a n d 5.25"diskettes in either MS-DOS or Macintosh format may be submitted. Files created with the current or near-current revisions of the following word processing programs are acceptable:
IBM Displaywrite Executive Writer Leading Edge Microsoft Word Multimate Officewriter PC Write PFS: Write PFS: Professional
Q & A Write Samna Spellbinder Volkswriter WordPerfect WordStar WordStar 2000 Xywrite
Macintosh MacWrite Microsoft Word WordPerfect All diskettes must be accompanied by the final revision of the manuscript and a duplicate copy, including two clearly separated and labeled sets of illustrations. All materials must be sent to the Editorial Office in St. Louis. Before submitting your diskettes, make certain to comply with the Guidelines for Electronic Manuscripts in the Information for Authors section of this journal. Should you have additional questions, please communicate with the Editorial Office: telephone (314) 361-6084; FAX (314) 367-0585.
Chest Wall Involvement by Lung Cancer: Computed Tomographic Detection and Results of Operation G. B. Ratto, MD, G. Piacenza, MD, C. Frola, MD, F. Musante, MD, I. Serrano, MD, R. Giua, MD, M. Salio, MD, P. Jacovoni, MD, and S. Rovida, MD Institute of Clinica Chimrgica, University of Genoa, Genoa, Italy
The aim of this prospective study was to evaluate: (1)the role of computed tomographic scanning in predicting chest wall invasion by peripheral lung cancer and (2) the results of operation according to the depth of chest wall involvement and other potential indicators of long-term survival. One hundred twelve patients with non-small cell lung cancer adjacent to the pleural surface who underwent computed tomographic scanning and subsequent thoracotomy were entered into this study. Tumor invasion was confined to the visceral pleura in 53 patients, to the parietal pleura in 18 patients, and to intercostal muscles in 25 patients; invasion extended beyond this layer in 16 patients. The computed tomographic criteria for chest wall invasion were (1)obliteration of the extrapleural fat plane, (2) the length of the
tumor-pleura contact, (3) the ratio between the tumorpleura contact and the tumor diameter, (4) the angle of the tumor with the pleura, (5) a mass involving the chest wall, and (6)rib destruction. The computed tomographic criteria 1 and 3 were significantly related to pathologic findings. Sensitivity was 85% for criterion 1 and 83% for criterion 3, specificity being 87% and 80%, respectively. Long-term survival of patients with T3 disease critically depended on the lymph node state and completeness of resection. The adenocarcinoma cell type and the T4 category were unfavorable prognostic factors. The depth of chest wall invasion did not affect survival, except for extensive rib and soft tissue infiltration. En bloc resection yielded better results than discontinuous resection. (Ann Thoruc Surg 2992;52:282-8)
S
Patients and Methods
urgeons have now recognized that patients with nonsmall cell lung cancer invading the chest wall have a potential long-term survival, provided that complete resection is feasible and mediastinal node involvement is not present [l-31. The decision whether to undertake en bloc resection of the lung and chest wall or to attempt extrapleural dissection depends on the presumed degree of peripheral tumor spread. Even intraoperatively, differentiation between neoplasms confined within the parietal pleura from those infiltrating the chest wall proper may be difficult. Preoperative assessment of chest wall involvement is therefore important in planning treatment. Although computed tomographic (CT) scanning has been suggested to be less than adequate to establish the depth of chest wall infiltration (4-61, it remains the most accurate diagnostic procedure currently available in Italy. The present prospective study was carried out to evaluate (1) the diagnostic accuracy of individual CT criteria and their combinations in predicting the extent of chest wall invasion and (2) the results of operation according to the depth of chest wall involvement and other potential indicators of long-term success.
Accepted for publication Sep 26, 1990. Address reprint requests to Dr Ratto, Istituto di Clinica Chirurgica la University of Genoa, Viale Benedetto XV, 10, 16132 Genoa, Italy.
0 1991 by The
Society of Thoracic Surgeons
From January 1983 through December 1988, patients admitted to four institutions (Clinica Chirurgica Ia, University of Genoa; San Martino and Sanpierdarena Hospitals, Genoa; and Alessandria Hospital, Alessandria) with peripheral, potentially resectable lung cancer were entered into this prospective study. The inclusion criteria were (1) histologic diagnosis of non-small cell lung cancer, (2) preoperative CT scanning demonstrating contiguity of the tumor with the pleural surface, and (3) thoracotomy performed. Subjects who had superior sulcus neoplasms, malignant pleural effusions, or preoperative radiotherapy and those who were not available to follow-up were excluded from the study. Computed tomographic images were read independently by two of us (C.F. and F.M.) before operation. They were asked if the tumor was confined within the parietal pleura or invaded extrapleural soft tissues. Preoperative staging procedures included chest roentgenography, bronchoscopy, and computed tomography of the chest and abdomen. Bone and brain scans, as well as cervical mediastinoscopy, were not included in the staging protocol. Throughout the entire study duration, the follow-up of patients was carried out according to Ginsberg’s suggestions [7]. There were 102 men and 10 women. The mean age was 8.6 years. Histology of the primary tumor was 59.8 squamous cell carcinoma (47 patients), adenocarcinoma (56), large cell carcinoma (5), and adenosquamous carci-
*
0003-4975/91/$3.50
Ann Thorac Surg 1991;Sl:182-8
noma (4). At operation, 13 neoplasms proved to be unresectable because of invasion of the spine, unremovable node metastases, or aortic arch infiltration. In 13 patients the resection was classified as incomplete because microscopic examination showed either the margins of the resected specimen or the most distally removed node in major lymphatic drainage regions to be infiltrated by tumor cells. Of the 86 subjects who underwent a complete resection, 21 had pneumonectomy and 65 lobectomy or bilobectomy. Biopsies of attainable paratracheal, subcarinal, inferior pulmonary ligament, and hilar nodes were routinely performed. The same surgical policy was applied by the surgeons participating in the study (G.B.R., I.S., R.G., and P.J.). The decision to perform an extrapleural dissection or en bloc resection depended on the operative findings. A trial of extrapleural dissection was attempted when there was no clear evidence of tumor extension into the intercostal muscles or ribs. This approach was used in 76 patients: 18 had parietal pleura invasion and 19 had tumor extension beyond the parietal pleura. In these latter patients, the involved chest wall was discontinuously resected. If during extrapleural dissection there was strong fixation of the parietal pleura to the overlying ribs, en bloc resection was carried out (18 patients). The number of resected ribs ranged from one to four. Prosthetic materials were never used for parietal reconstruction. Nineteen neoplasms were postsurgically classified as stage I, 9 as stage 11, 66 as stage IIIa, and 18 as stage IIIb. Histologic examination of the removed specimens demonstrated tumor involvement confined to the visceral pleura in 53 patients, parietal pleura involvement in 18, and intercostal muscle or rib periostium involvement in 25. In 16 patients there was either soft tissue infiltration beyond the intercostal muscles or rib destruction. Postoperative radiation therapy was given to patients with known residual tumor. Two patients died within 30 days after operation because of respiratory failure and intrathoracic bleeding. No death followed en bloc or discontinuous chest wall resection.
Computed Tomographic lnvestigations Computed tomographic scans of the chest were obtained using a Somatom I1 (Siemens, Erlangen, Germany) scanner. Contiguous cuts, 8 mm thick, were taken at full inspiration from the diaphragm to the lung apexes without overlap. Contrast medium was infused intravenously. The CT scanning criteria used to predict the depth of chest wall involvement were (1)obliteration of the extrapleural fat plane, (2) the length of the tumor contact with the adjacent pleural surface, (3) the ratio between the amount of tumor-pleura contact and the tumor diameter, (4) the angle of the tumor with the pleura (acute or obtuse angle), (5) an obvious mass extending into the chest wall soft tissues, and (6) rib destruction. Thickening of the pleura adjacent to the neoplasm was not taken into account because it has been previously demonstrated to be nonspecific [4]. The extrapleural fat plane was not uniformly discernable, particularly on CT cuts corresponding to the rib level. Nevertheless, the plane was practically always
RA'ITOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
183
Fig I. Right bronchogenic carcinoma contiguous with the chest wall. The extrapleural fat plane is well visible, with the arrow showing encroachment of tumor on extrapleural fat. Irregularity of the rib border, suggesting bone erosion, is also demonstrated.
visible if researched between the pleura adjacent to the tumor (which was often thickened) and the intercostal muscle layer (Fig 1).Measurement of tumor diameterwas obtained from CT scans using the soft-tissue window settings [4].Differentiation between cancer and surrounding atelectasis or pneumonitis was attempted by ahalyzing the vascularization patterns after contrast medium injection. The contact between abnormal lung and chest wall was assessed as shown in Figure 2. Rib destruction was searched for using the thin-cut technique and was defined as any evidence of irregular rib contour (see Fig 1). When the angle of the tumor with the pleura was either acute or obtuse in different areas, it was classified as obtuse.
Statistical Analysis Computed tomographic findings were correlated with pathologic findings, which were assumed as the true reference. Comparison was made in terms of sensitivity, specificity, and accuracy according to standard formulas. The predictive value of CT results was assessed both for each CT criterion and for their combinations. When a combination of several CT criteria was used, the diagnostic efficacy of the test was evaluated in two ways: (1) by classifying the test positive if at least one CT criterion indicated tumor infiltration beyond the parietal pleura, and (2) by considering the test positive if all the selected CT criteria suggested chest wall invasion. The strength of the association between the results provided by CT scanning and by thoracotomy was evaluated by multiple regression analysis. Because CT criteria 1, 5, and 6 were
184
RATTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Ann Thorac Surg 1991;51:182-8
Table 2. Diagnostic Efficacy of Computed Tomographic Criterion 2 (Length of Tumor Contact With Pleural Surface) According to Different Cut-off Values for Separating Patients With Chest Wall Invasion From Those Without Tumor-Pleura Contact (cm) Value Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (%)
Fig 2. L-$t bronchogenic carcinoma infiltrating the chest wall. The technique used for measuring the length of tumor in contact with pleura is illustrated. subjective in nature, intraobserver and interobserver variation have been estimated by means of the 4 coefficient method. This coefficient was 0.03 for intraobserver variation and 0.05 for interobserver variation. Survival rates have been calculated by the Kaplan-Meier method. Comparison of survival curves was done by the MantelHaensel test.
Results Sensitivity, specificity, and accuracy of individual CT criteria for the diagnosis of chest wall invasion, as well as the predictive value of positive or negative tests, are reported in Table 1. Sensitivity, specificity, and accuracy rates of CT criterion 2 (the length of the tumor contact with the adjacent pleural surface) when using different cut-off values separating patients with chest wall infiltration from those without infiltration are listed in Table 2. Sensitivity, specificity, and accuracy rates of CT criterion 3 Table 1. Computed Tomographic Scan Criteria for Chest Wall Invasion
Value Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (%)
Obliteration of the Chest Extrapleural Obtuse Wall Rib Fat Plane Angle Mass Destruction 85 87 86 80
57 72 66 55
33 100 75 100
16 98 67 87
91
73
71
66
97 42 61 51
92 54 67 54
76 68 71 59
97
93
83
(the ratio between the length of tumor-pleura contact and the tumor diameter) when adopting various cut-off values are summarized in Table 3. The correlation between CT data and pathologic findings was significant for CT criteria 1 ( p < 0.001) and 3 ( p < 0.01). Sensitivity, specificity, and accuracy rates of combining CT criteria 1 and 3 are shown in Table 4. The patients’ age and sex, the tumor diameter, and the number of resected ribs did not significantly affect survival. Survival rates depended on the tumor stage ( p < 0.05) and histology ( p < 0.05), the length of tumor-pleura contact ( p < 0.05), and the completeness of resection ( p < 0.001). The 5-year survival rate, according to the disease stage, was 48% in stage I, 31% in stage 11, 12% in stage IIIa, and 0% in stage IIIb. Survival curves of patients with T3 tumors, according to the N status, are illustrated in Figure 3. Patients with cancers classified as T3 NO survived longer than those with cancers classified as T3 N2 ( p < 0.05); patients with T3 N1 lesions also survived longer than those with T3 N2 lesions ( p < 0.05). Similar conclusions were obtained when the analysis was restricted to the individuals who had undergone a complete resection. The 5-year survival rate, according to the tumor cell type, was 25% in subjects with squamous cell carcinoma and Table 3 . Diagnostic Efficacy of Computed Tomographic Criterion 3 (Ratio Between Tumor-Pleura Contact and Tumor Diameter) According to Different Cut-off Values for Separating Patients With Chest Wall Invasion From Those Without Tumor-Pleura ContactTumor Diameter Value
0.5
0.7
0.9
Sensitivity (%) Specificity (%) Accuracy (%) Positive predictive value (%) Negative predictive value (a)
100
37 60 48
92 62 74 60
83 80 81 71
100
94
89
1991;51:182-8
Table 4 . Combined Use of Computed Tomographic Criteria 1 and 3 (Adopting 0.7 as Cut-of Value) for Predicting Chest Wall lnvasion Value
A
B
Sensitivity (%) Specificity (%)
97 61 75 60
81 89 86 81
98
89
Accuracy (%)
Positive predictive value (%) Negative predictive value (%)
185
RA'ITOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Ann Thorac Surg
100
90 80
70 60 50
A = test classified as positive if at least one criterion indicates chest wall invasion; B = test classified as positive if both criteria indicate chest wall invasion.
2o 10
17% in those with adenocarcinoma. When limiting the study to patients who had a potentially complete resection, the relevant figures were 30% and 19%, respectively. The length of the tumor-pleura contact affected survival exclusively when lesions adhering less than 3 cm were compared with lesions adhering more than 6 cm (either if considering the entire series or just the completely resected cases). The 5-year survival rate of patients with T3 tumors who underwent a complete resection (45 patients) was 16%. There were no 2-year survivors in the group of patients with T3 tumors who had incomplete resection or no resection (10 patients). Survival rates of patients with tumors extending to or beyond the parietal pleura who had undergone an extrapleural dissection, en bloc resection, or discontinuous resection of the involved lung and chest wall are shown in Figure 4. There was no significant difference in survival between patients of the extrapleural dissection group and
1 6
12 18 24
30 36 42 48 54 60
MONTHS
Fig 4 . Survival curves of patients with tumors extending to or beyond extrathe parietal pleura, according to the type of resection. (U en bloc resection, 18 patients; pleural dissection, 18 patients; 0-0 discontinuous resection, 19 patients.)
*-*
those of the en bloc resection group. Survival rates after a discontinuous resection were significantly poorer than those after en bloc resection ( p < 0.01). Survival rates according to the depth of chest wall infiltration are illustrated in Figure 5. The probability of long-term survival for patients with cancers confined to the visceral pleura was significantly greater than that of patients with rib destruction or soft tissue infiltration beyond the intercos-
x 100 90 80
70
60 50 40
30 20
\
10
1°1 O
6 12 18 24 30 36 42 48 54 60
HONTHS
Fig 5 . Survival curves of patients with tumors contiguous with the pleural surface, according to the depth of chest wall infiltration. 6 12 18 24 30 36 42 48 54 60 MONTHS (M tumor involvement confined within the visceral pleura, 53 patients; Hparietal pleural involvement, 18 patients; Fig 3 . Survival curves of patients with tumors extending to or beyond the parietal pleura (T3 tumors), according to the N status. (U intercostal muscle or rib periostial involvement, 25 patients; rib destruction or chest wall infiltration beyond intercostal muscles, 16 T3 NO, 14 patients; UT3 N1, 19 patients; T3 N2, 22 patients.) patients.)
0
*---*
*-*
*-*
186
RATTO ET AL CHEST WALL INVOLVEMENT BY LUNG CANCER
tal muscles ( p < 0.01 for the entire series; p < 0.05 for completely resected cases). No significant difference in survival was found among patient groups with intermediate degrees of chest wall infiltration.
Comment Preoperative diagnosis of chest wall invasion by primary lung cancer is important for surgical strategy. Mortality and morbidity associated with unnecessary extended resections [%lo] or unneeded violations of the tumor plane could be avoided. Combined therapy programs, including the use of preoperative radiotherapy, might be prospectively applied. Patients with chest wall invasion and mediastinal node metastases, who do not seem to benefit from en bloc resection [8, 10, 111, could be excluded from radical operations. Although CT scanning has been recognized as a more accurate diagnostic procedure than standard tomography [12-151, it has been recently suggested to be unsuitable to differentiate tumor contiguity with the pleural surface from definite chest wall invasion [ P 6 , 161. The individual criteria used to predict chest wall infiltration proved to be either highly sensitive but nonspecific or highly specific but insensitive. Among the criteria adopted in the present study, obliteration of the extrapleural fat plane was the most significantly related to pathologic findings. Both sensitivity (85%) and specificity (87%)rates were acceptable. The major criticism to the use of this criterion is the variable thickness of the extrapleural fat plane, often indistinguishable on CT images. In our experience, the fat plane between the pleura adjacent to or invaded by the tumor and the intercostal muscle layer was virtually always visible. One could object that, by focusing the analysis to the intercostal muscle regions, rib invasion without obvious bone destruction might have been overlooked. Actually, this occurred in 3 patients only, suggesting that the event is not common. Whether radionuclide bone scanning could be useful in identifying patients with rib involvement undetectable on CT scans remains to be clarified. A false diagnosis of extrapleural fat plane obliteration was made in 9 patients. According to Pennes and associates [4] and Shin and co-workers [17], the relevant causes have been identified in concomitant inflammatory processes of the pleura extending into the extrapleural fat or in chest asymmetry due to the patient’s position. Besides obliteration of the extrapleural fat plane, the only CT criterion significantly related to pathologic data was the ratio between the tumor-pleura contact and tumor diameter. One of the reasons accounting for this finding might reside in the fact that measurement errors were in part withdrawn by expressing the final results as a ratio. If surrounding atelectasis or pneumonitis would be erroneously deemed cancer, inaccuracy might affect the size of both the tumor diameter and tumor-pleura contact. All the remaining CT criteria that have been used in this study were not significantly associated with pathologic data. In an attempt to enhance the predictive value of CT findings, we combined the results of criteria 1 and 3. By using such an association, the following information
Ann Thorac Surg 1991;51:182-8
could be obtained: (1) if both criteria are negative, the surgeon may be almost certain that an extrapleural dissection will not violate the cancer plane (sensitivity: 97%); (2)if both criteria suggest chest wall invasion, the surgeon will be advised against extrapleural dissection (specificity: 89%). The relationship between the depth of chest wall infiltration and prognosis is one of the still-unsolved problems. Gronqvist and associates [18] and Mishina and colleagues [19] suggested that invasion of the pleura is a more favorable indicator than invasion of intercostal muscles, whereas rib involvement gives the worst prognosis. Neef [20] and Martini (211 claimed that patients with invasion of the chest wall proper have a poorer prognosis than patients with invasion of the parietal pleura only. By contrast, McCaughan and co-workers [l]found that the extent of chest wall involvement does not significantly affect survival. The present study showed that tumor infiltration confined within the visceral pleura is significantly associated with a better outcome if compared with infiltration of chest wall soft tissues beyond the intercostal muscles or rib destruction. No significant difference in survival was demonstrated between patients with parietal pleura involvement and those with involvement of the chest wall proper. The observation remained valid even when corrected for the completeness of resection and nodal metastases. However, it should be noted that patients with parietal pleura invasion or chest wall invasion were unevenly treated. The chosen operation in subjects with tumor growth confined to the parietal pleura was extrapleural dissection, whereas subjects with invasion beyond the parietal pleura underwent en bloc or discontinuous chest wall resection. Controversy continues concerning the value of potential indicators of long-term success in patients with lung cancer abutting the chest wall [l,2, 11,221. In accordance with previous reports [l, 10, 11, 191, the present study confirmed that long-term survival is highly dependent on the lymph node state and completeness of operation. The poor results we have obtained in patients with T3 N2 disease further stress the need for mediastinoscopic exploration in individuals whose CT scanning shows chest wall infiltration and enlarged mediastinal nodes [ll]. The lack, in our experience, of 2-year survivors after incomplete resection and radiotherapy does not support the efficacy of combining these therapeutic modalities. The other prognostic factors we found to be correlated with long-term survival were tumor histology, the T component, and the type of resection. Patients with squamous cell carcinoma had a better outcome than those with adenocarcinoma. This finding might suggest that the diagnostic approach to patients ’ with adenocarcinoma invading the chest wall should include a meticulous search for distant metastases to determine if they are actually MO. The erosion of survival expectation according to the T categories was significantly higher for T4 tumors than for T2 or T3 tumors. This would indicate that patients with vertebral body or great vessel invasion are poor candidates for surgical resection [2]. The prognostic value of the type of resection has been
Ann Thorac Surg 1991;51:182-8
differently appreciated by authors. Grillo [ll] found 3 long-term survivors in a group of 19 patients who had undergone en bloc resection, and none in a group of 14 patients in whom extrapleural dissection of a tumor involving the parietal pleura was done. Trastek and associates [23], analyzing subjects with lung cancer infiltrating the parietal pleura, noted 5-year survival rates of 75% after en bloc resection and 27.9% after extrapleural dissection. McCaughan and colleagues [ 11only attempted en bloc resection when there was operative evidence of intercostal muscle or rib invasion, and they concluded that survival is superior in patients undergoing extrapleural dissection (48%) than in those who had en bloc resection (16%). Casillas and co-workers [24] reported no significant difference in survival between the en bloc resection (21% at 5 years) and the extrapleural dissection (24%)groups. In the present series, the survival rates of patients who underwent en bloc resection were not statistically different from those of patients who underwent extrapleural dissection. Nevertheless, there was a tendency toward a better outcome in the first group than in the second group. This in spite of the more extended chest wall invasion characterizing subjects who required en bloc resection. We found survival to be significantly impaired when a discontinuous resection was carried out in place of en bloc resection. This observation is at variance with a previous report [lo] suggesting that discontinuous resection may be adequate in patients with tumor extension beyond the parietal pleura. The explanation might be that, in the earlier series, discontinuous resection was applied only when the adherence of the lung to the chest wall was neither solid nor extensive, whereas this was not always true in the present series. In conclusion, among the numerous CT criteria used for preoperative assessment of chest wall invasion by primary non-small cell lung cancer, obliteration of the extrapleural fat plane and the tumor-pleura contact-tumor diameter ratio were the only variables significantly related to pathologic findings. The depth of chest wall infiltration did not affect long-term survival, provided that extensive invasion of chest wall muscles or ribs was not present. Survival was critically dependent on the lymph node state and completeness of resection. The adenocarcinoma cell type and the T4 category were unfavorable prognostic factors. In patients with tumors extending beyond the parietal pleura, en bloc resection yielded better success rates than discontinuous chest wall resection. Whether en bloc resection is more advantageous than extrapleural dissection in patients with neoplasms confined t o the parietal pleura needs further investigation. The assistance of M. Galanti in the preparation of the manuscript is gratefully acknowledged.
References 1. McCaughan B, Martini N, Bains MS, McCormack PM. Chest
wall invasion in carcinoma of the lung. J Thorac Cardiovasc Surg 1985;89:83641.
RAlTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
187
2. Mountain CF. Value of the new TNM staging system for lung cancer. Chest 1989;96:47S-9S. 3. Miller KB, Little AG. Stage 111 lung cancer. Is surgery appropriate? Pulm Perspect 1989;6(2):2-5. 4. Pennes DR, Glazer GM, Wimbish KJ, Gross BH, Long RW, Orringer MB. Chest wall invasion by lung cancer: limitations of CT evaluation. AJR 1985;144:507-11. 5. Haggar AM, Pearlberg JL, Froelich JW, et al. Chest wall invasion by carcinoma of the lung. Detection by MR imaging. AJR 1987;148:1075-8. 6. Scott IR, Muller NL, Miller RR, Evans KG, Nelems 8. Resectable stage 111 lung cancer: CT, surgical, and pathologic correlation. Radiology 1988;166:75-9. 7. Ginsberg RJ. Follow-up supervision after resection for lung cancer. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery; vol 1: Lung cancer. Philadelphia: W.B. Saunders, 1985:274. 8. Piehler JM, Pairolero PC, Weiland LH, Offord KP, Payne WS, Bernatz PE. Bronchogenic carcinoma with chest wall invasion: factors affecting survival following en bloc resection. Ann Thorac Surg 1982;34:68&91. 9. Paone JF, Spees EK, Newton CG, Lillemoe KD, Kieffer RF, Gadacz TR. An appraisal of en bloc resection of peripheral bronchogenic carcinoma involving the thoracic wall. Chest 1982;81:203-7. 10. Patterson GA, Ilves R, Ginsberg RJ, Cooper JD, Todd TRJ, Pearson FG. The value of adjuvant radiotherapy in pulmonary and chest wall resection for bronchogenic carcinoma. Ann Thorac Surg 1982;34:692-7. 11. Grillo HC. Technical considerations in stage I11 disease: pleural and chest wall involvement. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery; vol 1: Lung cancer. Philadelphia: W.B. Saunders, 1985:134. 12. Ekholm S, Albrechtson U, Kugelberg J, Tylen U. Computed tomography in preoperative staging of bronchogenic carcinoma. J Comput Assist Tomogr 1980;4:76>5. 13. Baron LR, Levitt RG, Sagel SS, White MJ, Roper CL, Marbarger JP. Computed tomography in the preoperative evaluation of bronchogenic carcinoma. Radiology 1982;145: 727-32. 14. Epstein DM, Stephenson LW, Gefter WB, Van der Voorde F, Aronchik JM, Miller JT. Value of CT in the preoperative assessment of lung cancer: a survey of thoracic surgeons. Radiology 1986;161:423-7. 15. Rendina EA, Bognolo DA, Mineo TC, et al. Computed tomography for the evaluation of intrathoracic invasion by lung cancer. J Thorac Cardiovasc Surg 1987;94:57-63. 16. Glazer HS, Duncan-Meyer J, Aronberg DJ, Moran JF, Levitt RG, Sagel SS. Pleural and chest wall invasion in bronchogenic carcinoma: CT evaluation. Radiology 1985;157 1914 17. Shin MS, Anderson SD, Myers J, Ho KJ. Pitfalls in CT evaluation of chest wall invasion by lung cancer. J Comput Assist Tomogr 1986;10136-8. 18. Gronqvist YKJ, Clagett OT, MacDonald JR. Involvement of thoracic wall in bronchogenic carcinoma: study of 16 cases in which pneumonectomy or lobectomy and simultaneous resection of the thoracic wall were done. J Thorac Surg 1957; 33:487-95. 19. Mishina H, Suemasu K, Yoneyama T. Surgical pathology and prognosis of the combined resection of chest wall and lung in lung cancer. Jpn J Clin Oncol 1978;8:161-8. 20. Neef H. Discussion of CasiIlas M, Pans F, Tarrazona V,
188
RAlTOETAL CHEST WALL INVOLVEMENT BY LUNG CANCER
Padilla J, Paniagua M, Galan G. Surgical treatment of lung carcinoma involving the chest wall. Eur J Cardiothorac Surg 1989;3:425-9. 21. Martini N. Discussion of: Patterson GA, Ilves R, Ginsberg RJ, Cooper JD, Todd TRJ,Pearson FG. The value of adjuvant radiotherapy in pulmonary and chest wall resection for bronchogenic carcinoma. Ann Thorac Surg 1982;34:692-7. 22. Sellman M, Heme A, Peterffy A. Extended intrathoracic
Ann Thorac Surg 1991;51:182-8
resection for lung cancer. Scand J Thorac Cardiovasc Surg 1987;21:69-72. 23. Trastek VF, Pairolero PC, Piehler JM, et al. En bloc (non-chest wall) resection for bronchogenic carcinoma with parietal fixation. J Thorac Cardiovasc Surg 1984;87352-8. 24. Casillas M, Paris F, Tarrazona V, Padilla J, Paniagua M, Galan G. Surgical treatment of lung carcinoma involving the chest wall. Eur J Cardiothorac Surg 1989;3:425-9.
IrnDortant Announcement for Authors Elsevier Science Publishing Company has recently instituted electronic publishing for certain of their journals, including The Annals of Thoracic Surgery. Beginning September 1, 1990, the final version of accepted manuscripts submitted on diskettes will be converted electronically directly to page proof, with a considerable saving of time a n d expense. Authors are encouraged to use this option if they wish to do so. Both 3.5"a n d 5.25"diskettes in either MS-DOS or Macintosh format may be submitted. Files created with the current or near-current revisions of the following word processing programs are acceptable:
IBM Displaywrite Executive Writer Leading Edge Microsoft Word Multimate Officewriter PC Write PFS: Write PFS: Professional
Q & A Write Samna Spellbinder Volkswriter WordPerfect WordStar WordStar 2000 Xywrite
Macintosh MacWrite Microsoft Word WordPerfect All diskettes must be accompanied by the final revision of the manuscript and a duplicate copy, including two clearly separated and labeled sets of illustrations. All materials must be sent to the Editorial Office in St. Louis. Before submitting your diskettes, make certain to comply with the Guidelines for Electronic Manuscripts in the Information for Authors section of this journal. Should you have additional questions, please communicate with the Editorial Office: telephone (314) 361-6084; FAX (314) 367-0585.
