Invasion to the visceral pleura is an important component

Diagnosis of Visceral Pleural Invasion by Lung Cancer Using Intraoperative Touch Cytology Yushi Saito, MD, PhD, Yosuke Yamakawa, MD, PhD, Masanobu Kiriyama, MD, PhD, Ichiro Fukai, MD, PhD, Satoshi Kondo, MD, PhD, Masahiro Kaji, MD, Motoki Yano, MD, Tomoki Yokoyama, MD, and Yoshitaka Fujii, MD, PhD Department of Surgery II, Nagoya City University Medical School, Nagoya, Japan Background. Invasion to the visceral pleura is an important component of lung cancer staging and an independent prognostic factor. However, the accuracy of pathologic examination depends on how the sections are made, and the pathologist may miss the most invaded part of the pleura. Therefore, we have designed touch cytology in an effort to more accurately diagnose the pleural invasion by lung cancer. Methods. Immediately after thoracotomy, the surface of the visceral pleura just above the tumor was gently touched by a glass slide without scrubbing in 100 patients who simultaneously underwent pleural lavage cytology or cytology of the subclinical pleural effusion. Results. Seventeen percent of the tumors were diagnosed as invading the visceral pleura by touch cytology. Lavage cytology was found to be positive in 7%. In reference to the pathologic examination of the tumor specimen, touch cytology was found to be positive in all of p3, 5 out of 6 of p2, 5 out of 30 of p1, and 5 out of 62 of p0 cases. correctly diagnosed all the positive cases detected by lavage or effusion cytology. Conclusions. This study suggests that our method is useful in detecting the visceral pleural invasion and raises a possibility that pathologic p0 and p1 lung cancers include a subset of patients with tumor cells exposed on the pleural surface. (Ann Thorac Surg 2002;73:1552 7) 2002 by The Society of Thoracic Surgeons Invasion to the visceral pleura is an important component of lung cancer staging and an independent prognostic factor [1 3]. The presence of pleural invasion in carcinoma smaller than 3 cm increases the staging descriptor T1 to T2 [4]. Significant differences in survival have been recorded between the T1 and T2 disease [5 6]. In most cases, the visceral pleura in one or two sections are examined by the pathologist for the diagnosis of pleural invasion. Pathologic examination is thus limited by the way the sample is cut from the specimen and the pathologist may fail to examine the most invaded part of the pleura. Lavage cytology is a good complementation to the diagnosis of pleural invasion. Once the cancer cells are released to the chest cavity from the tumor that invaded the pleura, there is a good chance of detecting the malignant cells by the lavage cytology. However, the sensitivity of this method may be low if the invasion reaches the surface of the pleura but releases few cancer cells. We have designed touch cytology in which we gently touch the pleura overlying the tumor with a piece of glass slide. We hoped that this touch method may improve the sensitivity of detecting the pleural invasion by lung cancer. In this article we report 100 patients of lung cancer who underwent touch cytology together with lavage cytology or cytology of the subclinical pleural effusion. Accepted for publication Dec 30, 2001. Address reprint requests to Dr Fujii, Department of Surgery II, Nagoya City University Medical School, 1, Kawasumi, Mizuhoku, Nagoya 467-8601, Japan; e-mail: yosfujii@med.nagoya-cu.ac.jp. Patients and Methods From July 1998 to September 2000, 175 patients underwent surgical resection at Nagoya City University Medical School. Of these, 108 patients were available for evaluation of touch cytology of the visceral pleura. Among these, 100 patients also underwent pleural lavage cytology or cytology of the pleural effusion, which was not apparent before the operation. These 100 patients were the subjects of the present study. Patients with apparent pleural effusion preoperatively were excluded. There were 73 adenocarcinomas, 21 squamous cell carcinomas, five adenosquamous carcinomas and one small cell carcinoma. Preoperatively 18 patients had undetected pleural effusion; the effusion was subjected to cytology. Ninety-four patients underwent lavage cytology. Twelve patients underwent both effusion and lavage cytology. Our method of touch cytology is as follows. Immediately after thoracotomy, the surface of the visceral pleura just above the tumor is gently touched by a glass slide without scrubbing (Fig 1). This method is repeated 4 to 5 times using a new piece of glass slide each time. Subsequently, 100 ml of saline solution is instilled into the chest cavity and the washing is aspirated for lavage cytology. There is no manipulation of the lung before or during this procedure. In our preliminary studies, we checked the section of the touched lung and confirmed that touching the pleura by a piece of glass did not affect the subsequent pathologic examination of the pleura and lung. 2002 by The Society of Thoracic Surgeons 0003-4975/02/$22.00 Published by Elsevier Science Inc PII S0003-4975(02)03404-5

Ann Thorac Surg SAITO ET AL 2002;73:1552 7 TOUCH CYTOLOGY OF VISCERAL PLEURA 1553 Fig 1. The technique of touch cytology was used to detect visceral pleural invasion of a tumor. Immediately after thoracotomy, the pleura overlying the tumor was gently touched by a piece of glass slide. This was repeated five times. Care was taken not to scrub the surface. The obtained materials were stained by Giemsa and Papanicolaou s method for cytologic examination. Immunocytochemical staining with CEA and MOC-31 was also performed to distinguish reactive mesothelial cells from cancer cells. For the routine histologic diagnosis of pleural invasion, parts of the tumor that appeared to have macroscopic visceral pleural invasion were sliced. In most cases the visceral pleura of one or two cut slices of the resected tumor were examined by one pathologist. Elastic fiber was stained using Victoria Blue to identify the pleura. Visceral pleural involvement was classified according to the rules of the Japan Lung Cancer Society [7] as follows: p0 no visceral pleural involvement; p1 penetration through the visceral pleura elastic tissue but not to the surface of the visceral pleura; p2 penetration through to the surface of the visceral pleura; and p3 involvement of the parietal pleura. Chi-square test was used to determine the significance of the relationship between the two variables. A p value of less than 0.05 was considered significant. Results Of the 100 tumors, 17 tumors were diagnosed as invading the visceral pleura by touch cytology. Lavage cytology detected only 7 of these (7% positive). Cytology of the pleural effusion, which was found at the time of operation, was positive in only 3 of 18 patients (Table 1). The frequency of positive touch cytology in patients with adenocarcinoma, squamous cell carcinoma, adenosquamous carcinoma, and small cell carcinoma was 21% (15 of 73), 4.8% (1 of 21), 0% (0 of 5), and 100% (1 of 1), respectively (Table 1). The result of touch cytology was analyzed in reference to the pathologic examination of the tumor specimen (Table 2). For two pathologic p3 tumors, touch cytology correctly identified all of them. However, it failed to detect one of six p2 patients. To our surprise, among the tumors that were pathologically noninvading to the pleural surface (p0 and p1), touch cytology was found to be positive in 5 of 62 (8.1%) p0 and in 5 of 30 (16.7%) p1 patients. The number of tumor cell clusters on the touch cytology slides tended to be higher in more pathologic invading tumors (p 0.17; Table 2). The nodal status of the tumor did not have a statistical significant impact on the result of touch cytology (p 0.62). However, n1 and n2 patients tended to have more cancer cell clusters (3 of 3 n1 and 4 of 6 n2 patients had 5 or more clusters) than n1 patients (1 of 8 who had 5 or more clusters, p 0.17). Hemotoxylin and eosin stains of the selected sections and the corresponding touch cytology slides are shown in pairs in Figure 2, which illustrates the pitfalls of both pathologic examination and touch cytology. Figures 2A and 2B show a patient with pleural invasion (p2) by hemotoxylin and eosin histologic examination. In this patient touch cytology also shows clusters of tumor cells among mesothelial cells. Figures 2C and 2D illustrate a patient in which pathologic examination failed to detect tumor cells exposed to the pleural surface (p1), but tumor cells were found on touch cytology. In this patient the tumor extended beyond the elastic fibrous layer of the visceral pleura, but the tumor cells were judged as not being exposed on the pleural surface because of the thick desmoplastic fibrous tissue overlying the visceral pleura. In this patient touch cytology of the visceral pleura Table 1. Rate of Visceral Pleural Invasion Diagnosed by Touch Cytology, Effusion, and Lavage Cytology in Reference to Histological Subtype Histological Subtype Total Adenocarcinoma Squamous Cell Carcinoma Adenosquamous Carcinoma Small Cell Carcinoma Positive ratio 17/100 (17%) Lavage cytology Positive ratio 7/94 (7%) Effusion cytology Positive ratio 3/18 (17%) 15/73 (21%) 6/71 (8%) 3/10 (30%) 1/21 (5%) 0/19 0/6 0/5 0/3 0/2 1/1 (100%) 1/1 (100%) Not done...

Ann Thorac Surg SAITO ET AL 2002;73:1552 7 TOUCH CYTOLOGY OF VISCERAL PLEURA 1555 Table 2. Rate of Visceral Pleural Invasion Diagnosed by Touch Cytology in Reference to the Pathologic Examination Pathologic Examination p0 p1 p2 p3 Positive/total 5/62 5/30 5/6 2/2 Numbers of cancer cell clusters 1 3 4 1 2 2 5 or more 1 4 3 0 Table 3. Correlation Between the Result of Touch Cytology and Pleural Effusion or Lavage Cytology Effusion Cytology Lavage Cytology Negative Positive Negative Positive Negative (n 83) 14 0 78 0 Positive (n 17) 1 3 9 7 Number of cancer cell clusters 1 3 1 0 8 1 5 or more 0 3 1 6 showed many clusters of malignant cells. Figures 2E and 2F show a patient with no pleural involvement by elastic stain (p0). A few malignant cells were found in the touch cytology specimens among mesothelial cells. Figure 2G illustrates a difficult patient for touch cytology; the tumor with visceral pleural invasion is visible at the bottom of a deep groove made by the pleural retraction common to an adenocarcinoma. No tumor cells could be identified by touch cytology. Lavage cytology was also negative in this patient. Table 3 shows the correlation between the result of touch cytology and those of pleural effusion or lavage cytology. correctly diagnosed all the positive patients detected by lavage or effusion cytology. In 8 of 17 patients with positive touch cytology, neither effusion or lavage cytology could detect malignant cells. Lavage cytology detected only 7 of 16 patients (44%) with positive touch cytology. One patient had pleural effusion at the time of operation that was negative and the lavage cytology was also negative. In patients with many tumor cell clusters on touch cytology, the chance of positive effusion or cytology was high (90%). Comment In this article we described touch cytology as gently touching the pleura overlying the tumor in an effort to more accurately diagnose the pleural invasion by lung cancer. Using touch cytology we have shown that some pathologic p0 and p1 patients actually may have had tumor cells that were easily detachable from the pleural surface. also was found to be superior to the conventional pleural lavage cytology. We found positive touch cytology in 5 of 62 (8.1%) and 5 of 30 (16.7%) pathologic p0 and p1 patients, respectively. This may be attributed to inappropriate preparation of sections for pathologic examination, and there may actually have been tumor cells exposed on the pleural surface. In this case pathologic examination failed to detect about 11% of true positives. Alternatively it could be argued that by touching the pleura we may have been damaging the pleura by exposing the otherwise unexposed tumor cells. However, this is unlikely because the pathologist did not observe damaged pleura in any of the patients. Among 62 pathologic p0 cases, touch cytology showed positive findings in 5 patients. It is rather difficult to imagine that at one section the tumor does not even reach the elastic fibrous sheet of the pleura and at another region of the same pleura it may be exposed on the surface. Kondo and colleagues [8] suggested that the exfoliation of cancer cells into the pleural cavity may occur not only when the tumor is exposed on the pleural surface but also when subpleural lymphatics are invaded by the tumor. However, we could not demonstrate lymphatic spread in our series of patients with positive touch cytology and a pathologic diagnosis of p0. Because morphologic distinction between reactive pleural mesothelial cells and adenocarcinoma cells could be difficult, the accuracy of cytologic examination remains to be one of the problems to be solved. In patients like the one shown in Figure 2C (in which pathologic p1 tumors that invade the elastic fiber but are retained under the pleural surface), it is conceivable that touch cytology can detect tumor cells on the surface, whereas in some sections, tumor cells are retained within the confinement of the fibrous tissue of the pleural surface. In many p1 patients, touch cytology revealed 5 or more clusters of tumor cells, which strongly suggests truly positive tumor cells on the pleural surface. In only 1 patient of pathologic p2 lung cancer, touch cytology failed to show tumor cells on the slide that touched the pleura overlying the tumor (Fig 2G); this was 4 Fig 2. Pathologic findings of representative patients. Hematoxylin & eosin (HE) staining: (A), (C), (E), and (G). Elastic fiber staining (blue): (C), (E), and (G). : (B), (D), and (F). (A) and (B) show a p2 patient. (A) HE staining shows lung cancer invading the pleura (arrow). (B) shows tumor cells (arrow) among mesothelial cells. (C) and (D) show a p1 patient. In (C) the tumor cells apparently invade the elastic lamina (arrow), but they do not seem to be exposed to the surface, which is covered by a thick fibrous tissue thus leading to a diagnosis of p1. In this patient, touch cytology shows many clusters of malignant cells (arrow) as shown in (D). (E) and (F) show a p0 patient with no pleural involvement as shown by the blue elastic stain, (E). However, in (F) a few malignant cells (arrow) are found in the touch cytology specimens among mesothelial cells. (G) A case with pleural invasion (arrow) at the bottom of a deep groove made by the pleural retraction. No tumor cells could be identified by touch cytology in this patient (not shown).

1556 SAITO ET AL Ann Thorac Surg TOUCH CYTOLOGY OF VISCERAL PLEURA 2002;73:1552 7 one example of adenocarcinoma that had pleural indentation and tumor cells that were found to be exposed at the bottom of the deep crevice formed by the retracted pleura. This may have been a patient that was p0 in practice. The lavage cytology was also negative in this patient. In another 7 p2 and p3 patients, touch cytology could correctly identify the tumor cells (Table 2). Lavage cytology of the pleural space proved valuable in predicting the outcome of lung cancer patients. Okumura and colleagues [9] reported a 17.4% recurrence rate of p1 lung cancers with positive lavage cytology compared with 0.7% with negative lavage cytology. Kondo and coworkers [8] and Dresler and coworkers [10] reported 3-year survival rates of lavage cytology positive and negative stage I lung cancers to be 68.7% and 22.9% (Kondo and colleagues [8]) and 69% and 0% (Dresler and colleagues [10]), respectively. Other methods for more accurate detection of visceral pleural invasion have been proposed. Ichinose developed a method using a jet stream of saline solution [11]. These studies have suggested that a significant number of patients with otherwise early stage disease will have positive pleural cytology results and subsequently have a poorer prognosis than patients with negative cytology results. In our study, the touch cytology results had a good correlation with effusion or lavage cytology results (Table 3). Cases with negative touch cytology were always negative with effusion or lavage cytology. Without compromising specificity, touch cytology achieved better sensitivity than the effusion or lavage cytology. Our study raises a possibility that pathologic p0 and p1 lung cancers include a subset of patients with tumor cells exposed on the pleural surface. We have to wait for the follow-up study of these patients to see if the positive touch cytology has an impact on the survival of patients with otherwise early stage lung cancer. References 1. Ichinose Y, Yano T, Asoh H, Yokoyama H, Yoshino I, Katsuda Y. Prognostic factors obtained by a pathologic examination in completely resected non-small-cell lung cancer. J Thorac Cardiovasc Surg 1995;110:601 5. 2. Harpole DH, Herndon JE, Young WG, Wolfe WG, Sabiston DC. Stage I nonsmall cell lung cancer. Cancer 1995;76:787 96. 3. Inoue K, Sato M, Fujimura S, et al. Prognostic assessment of 1310 patients with non-small-cell lung cancer who underwent complete resection from 1980 1993. J Thorac Cardiovasc Surg 1998;116:407 11. 4. Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111:1486 7. 5. Bunker ML, Raab SS, Landreneau RJ, Silverman JF. The diagnosis and significance of visceral pleural invasion in lung carcinoma. Am J Clin Pathol 1999;112:777 83. 6. Barry G and Stefan JU. The significance of pleural elastica invasion by lung carcinomas. Hum Pathol 1990;21:512 7. 7. The Japan Lung Cancer Society. General rules for clinical and pathologic record of lung cancer, 5th ed. Tokyo: Kanehara, 1999. 8. Kondo H, Asamura H, Suemasu K, et al. Prognostic significance of pleural lavage cytology immediately after thoracotomy in patients with lung cancer. J Thorac Cardiovasc Surg 1993;106:1092 7. 9. Okumura M, Ohshima S, Kotake Y, Morino H, Kikui M, Yasumitsu T. Intraoperative pleural lavage cytology in lung cancer patients. Ann Thorac Surg 1991;51:599 604. 10. Dresler CM, Fratelli C, Babb J. Prognostic value of pleural lavage in patients with lung cancer resection. Ann Thorac Surg 1997;67:1435 9. 11. Ichinose Y, Yano T, Asoh H, et al. Diagnosis of visceral pleural invasion in resected lung cancer using a jet stream of saline solution. Ann Thorac Surg 1997;64:1626 9. INVITED COMMENTARY The authors describe a novel technique for assessing the visceral pleura during operations for lung cancer intraoperative touch cytology (ITC). Overall, ITC was positive in 17% of cases versus 7% for pleural lavage cytology (PLC). It is noteworthy that ITC was positive in a significant number of patients without pleural transgression detected by standard histology. There are several potential ramifications of these findings. The first question is whether this technique should supercede standard histology in classifying the T factor for staging lung cancer. Unless and until there is considerable further experience, the answer must be negative. If a surgeon is concerned that invasion has been missed by the pathologist, he or she should request further cuts. What we see grossly as visceral pleural puckering is not always pleural involvement, but may be a desmoplastic reaction to the underlying neoplasm. Second, the relationship of ITC and PLC cytology is obvious. Both techniques are based on the detection of malignant cells that have exfoliated from the primary tumor into the pleural space (lavage) or can be assumed to have a high potential to do so (touch). This preliminary study suggests that the touch prep may be more sensitive. Although PLC was described over 60 years ago, the literature on the subject is limited. Different techniques have been used with respect to the amount and type of fluid and whether lavage is done presection, postresection, or both. In addition, the available experience varies in regard to associated risk factors for positive lavage, such as T and N factor, as well as stage in general. Very importantly, however, it is clear that patients who have a positive lavage, other factors being equal, have a significantly worse long-term survival than those with negative cytology. Nonetheless, this finding does not at present affect TNM staging nor is there a standard approach to postoperative adjuvant treatment in this setting. It is clear from this study as well as those of PLC that during many operations for lung cancer, malignant cells are present in the pleural space and/or on the very surface of the lung. A personal practical approach to this fact is that I bathe the pleural space in distilled water after every resection in order to lyse cells. After resection, all operators change gloves and remove from the operative field all instruments or sponges that have contacted 2002 by The Society of Thoracic Surgeons 0003-4975/02/$22.00 Published by Elsevier Science Inc PII S0003-4975(02)03540-3