"Collagen balls" in peritoneal washings. Prevalence, morphology, origin and significance.

Peritoneal washings obtained at laparotomy from women undergoing surgery for neoplasms of the genital tract may contain "collagen balls," consisting of tissue fragments composed of collagen covered with mesothelial cells. Collagen balls were found in 19 (4.5%) of 418 peritoneal washings and were more prevalent in specimens labeled pelvic washings (17 of 294, or 5.8%) than in those labeled peritoneal washings (2 of 124, or 1.6%). In 15 of the 19 cases in which we found collagen balls, at least one ovary was available for microscopic examination. In 14 of the 15 cases minute nodular papillary stromal projections covered with mesothelium were found on the surface of the ovaries. We conclude that collagen balls, a nonspecific entity, most probably originate on the surface of the ovaries. Their significance lies in their being mistaken for mucin-distended cells exfoliated from a neoplasm or from detached fragments of a papillary ovarian neoplasm.     

Curschmann's spirals in pleural and peritoneal fluids. Report of 12 cases

A limited prospective study of 334 peritoneal and pleural fluids demonstrated Curschmann's spirals in 12 specimens, a prevalence of 1 in 28 cases. Nine of these 12 specimens were peritoneal washings, 2 specimens were pleural fluids, and 1 specimen was a peritoneal dialysis fluid. The fluids came from nine women and three men. Both pleural fluids were from men (one with a history of lung adenocarcinoma and one with probable collagen vascular disease). The peritoneal dialysis fluid was from a man with renal failure. Seven of the nine women with peritoneal washing specimens had gynecologic cancer; only one had evidence of serosal involvement by neoplasm at the time of the peritoneal washing. Tissue, including serosal surfaces, was available for examination in 10 of the 12 cases. The study indicates that Curschmann's spirals are a relatively common finding in peritoneal and pleural fluids, particularly in peritoneal washings. Mucin-producing epithelium or involvement of serosa by malignant neoplasm is not necessary for the phenomenon to take place. The feature common to all the cases in which tissue was available for review was the presence of myxoid degenerative changes of the serosal and subserosal fibrous connective tissue.     

Cytohistologic correlation of peritoneal washing cytology in gynecologic disease

The peritoneal washings and cul de sac aspirates from 204 patients undergoing 217 procedures for the evaluation of gynecologic disease were examined retrospectively and correlated with the histologic diagnoses. Of the 73 washings from patients with histologically benign genital disease, cytology diagnosed 64 (87.7%) as negative, 6 (8.2%) as inconclusive and 3 as malignant. One malignant washing was a true positive from a nongenital primary. False positives thus occurred in 2.7% of the benign cases on blind review. Of 144 cytologic examinations of washings from patients with histologically confirmed malignant disease of the female genital tract, 38 (26.4%) were considered positive after cytohistologic correlation. Four malignant cases (2.1%) were undercalled on blind review while 3 (2.0%) were considered overcalls. Eleven of 47 cases (23.4%) with biopsy-proven peritoneal disease had negative cytology after histologic correlation. Recurring problems in interpreting peritoneal washings included: (1) the differential diagnosis of the spectrum encompassed by reactive mesothelium, endosalpingiosis, borderline serous tumors and well-differentiated serous cystadenocarcinoma; (2) morphologic similarities between some tumor cells and mesothelial cells associated with treatment effects; and (3) a paucity of malignant cells in some washings, resulting in false-negative interpretations. Ineffective cytopreparation, particularly of bloody specimens, hampered interpretation of some specimens. Correlation with previous histology and cytology enhanced the accuracy of peritoneal washing cytology in this study.     

Collagen balls in cervical smears. A previously undescribed finding

OBJECTIVE: To evaluate several conventional cervical smears obtained from women undergoing routine screening for cervical dysplasia or carcinoma and whose smears contained structures resembling collagen balls. STUDY DESIGN: Between 1995 and 1998, cervical smears containing collagen balls were analyzed. The clinical histories of the patients whose smears contained collagen balls, including the gestational history, were reviewed. Histopathologic material from any related surgical specimens was reviewed, with special attention to mesothelial surfaces. RESULTS: Collagen balls were found in 5 of 77,891 Pap smears examined (0.006%). None of the patients had evidence of neoplasms of the genital tract. One of the patients was in the first trimester of pregnancy. CONCLUSION: We suggest that collagen balls in cervical smear originate in mesothelium-covered organs, from where they are transported via the fallopian tubes into the uterine cavity. Their significance lies in their being mistaken for mucin-distended cells exfoliated from a neoplasm or from detached fragments of a papillary ovarian neoplasm.     

Cytology of benign multicystic peritoneal mesothelioma in peritoneal washings

Objective:  To describe the cytological aspect of peritoneal washings in benign multicystic peritoneal mesothelioma (BMPM).
Methods:  Three peritoneal washing specimens stained by standard cytological and histological procedures and analysed by light microscopy.
Results:  The specimens showed an abundance of monomorphous mesothelial cells devoid of atypia or mitoses. The mesothelial cells were calretinin positive. They also showed numerous squamous metaplastic cells arranged in flat sheets or isolated cells. The background contained some inflammatory cells.
Conclusion:  The combination of cytology of the peritoneal washing, histology (cell block and surgical specimen) and clinical history allow differentiation of BMPM from other cystic lesions (cystic lymphangioma and malignant mesothelioma).     

Peritoneal washing cytology in cervical carcinoma. Analysis of 109 patients

The peritoneal washing cytologies of 109 patients (112 procedures) undergoing laparotomy for cervical carcinoma were evaluated retrospectively and compared with the clinical and pathologic findings. Nine patients (8.3%) had malignant peritoneal washings (including three of four with washings initially termed "inconclusive"). Four (4.9%) of the 82 patients with squamous carcinoma and 3 (16.7%) of 18 with adenocarcinoma had positive washings. Five (5.6%) of 90 washings obtained at initial explorations were positive, as compared with 4 (18.2%) of 22 washings obtained as follow-up operations in recurrent cases. The 111 peritoneal washing cytologies with a corresponding histologic evaluation of the peritoneal cavity showed a good correlation; peritoneal washing cytology had an efficiency of 91.0%, a sensitivity of 52.9% and a specificity of 100%. Two cases in which the cytologies were considered positive only after review had negative peritoneal histologies; both patients died of progressive disease within 11 months. Peritoneal washing cytology was positive in 5 (5.9%) of 84 cases with FIGO stage 1 cancers, 2 (18.2%) of 11 cases with stage 2 cancers, 1 (33.3%) of 3 cases with stage 3 cancers, and 1 (10%) of 10 cases with recurrent tumors. Eight (88.9%) of nine patients with malignant peritoneal washings died of disease from 3 to 15 months following surgery; one showed no evidence of disease at 9 months. These results suggest that: (1) cervical carcinomas are infrequently associated with a positive peritoneal washing; (2) peritoneal washing cytology is more likely to be positive in cases of adenocarcinoma than in cases of squamous carcinoma; (3) peritoneal washings obtained at the time of surgery for recurrence are more likely to contain malignant cells than are washings obtained during initial exploration; (4) nonkeratinizing malignant squamous cells may be confused with reactive mesothelial cells; and (5) peritoneal washing cytology is a relatively insensitive technique for detecting advanced cervical disease, but correlates with a poor prognosis when positive.     

Detection of adenocarcinoma in peritoneal washings by staining with monoclonal antibody B72.3

Peritoneal washings are routinely performed in the staging evaluation of carcinomas of the ovary and in "second-look" explorations; major problems in the evaluation of these specimens continue to be the distinction between atypical mesothelium and adenocarcinoma and the identification in an otherwise inflammatory specimen of rare cells of adenocarcinoma, which may be undetected by the most trained individual. Monoclonal antibody (MAb) B72.3, reactive with an oncofetal, tumor-associated glycoprotein (termed TAG-72; MW greater than 1000 kd) expressed in a variety of epithelial malignancies but not generally expressed in benign or malignant mesothelium, was reacted with sections from the paraffin-embedded cell blocks of 185 peritoneal washings from 180 patients with extant cancer or a prior history of malignancy. One hundred four of the washings were initially interpreted as atypical mesothelium, with no evidence of malignancy; when reacted with MAb B72.3, 6 of these specimens demonstrated groups of metastatic adenocarcinoma cells not appreciated by the usual cytologic criteria. Of the 81 washings interpreted as showing cells of adenocarcinoma, 73 demonstrated expression of TAG-72 from both gynecologic and nongynecologic malignancies. In the remaining 49 cases without an associated malignant process, MAb B72.3 did not stain atypical mesothelium, but did react, however, with benign endometrial cells and müllerian inclusions in two cases. MAb B72.3 may be used as a diagnostic adjunct to the routine cytologic evaluation of malignancy in peritoneal washings; reactivity with MAb B72.3 may indicate the need for further evaluation to define the presence of a malignancy or an advanced cancer.     

The value of calretinin and cytokeratin 5/6 as markers for mesothelioma in cell block preparations of serous effusions

Objective: To determine the value of calretinin and cytokeratin (CK) 5/6 in discriminating mesothelioma from adenocarcinoma in serous effusion specimens. Methods: A total of 101 recent, histologically or clinically confirmed malignant effusions with immunostained cell block preparations were reviewed. The cases consisted of 34 mesotheliomas and 67 adenocarcinomas. This included 17 ascitic fluid and 84 pleural fluid samples. The adenocarcinomas included metastatic carcinomas from the breast (12), lung (19), stomach (3), colon (1), pancreas (2), ovary (6) endometrium (1) and 23 histologically confirmed metastases from unknown primary sites. The cases were assessed as negative or positive (>5% of cells stained). The staining pattern was recorded as cytoplasmic, cell membrane, nuclear or cytoplasmic and nuclear staining. Results: Calretinin staining was present in 97% (33/34) of the mesothelioma cases with a majority of them showing both cytoplasmic and nuclear staining (29/33). Only 3% (2/67) of adenocarcinomas were positive for calretinin, one being a lung adenocarcinoma and the other an adenocarcinoma of unknown primary site in an ascitic fluid. Cytokeratin 5/6 staining was also present in 33/34 (97%) of mesothelioma cases. Six (9%) adenocarcinomas were positive, including metastases from the lung (1), breast (1), ovary (2) and unknown primary site (2). Four of the six adenocarcinoma cases positive for CK5/6 were in ascitic fluids. No cases of mesothelioma were negative for both calretinin and CK5/6. Only one adenocarcinoma case, (which was from unknown primary site in an ascitic fluid sample), was positive for both markers. Conclusions: The results confirm that calretinin and CK 5/6 are useful markers for mesothelioma in effusion specimens. CK5/6 staining may be less useful for peritoneal fluid specimens where metastatic adenocarcinomas may be more likely to express the antigen. Further study of ascitic/peritoneal specimens is warranted. However, positive staining, particularly for both antigens, is highly indicative of a mesothelial origin for cells. The two markers make a useful addition to EMA and the panel of adenocarcinoma markers routinely applied to effusion specimens.     

Collagen balls in cervica-vaginal smears

OBJECTIVE: To dewribe the recent detection of collagen balls in cervical-vaginal specimens from 7 women within the previous 2 years. STUDY DESIGN: Papaniolaou-stained cervical-vaginal specimens, 3 conventional and 4 ThinPreps (Cytyc Corp., Boxborgugh, Massachusetts, U.S.A.), were reviewed due to the presence of structures indistinguishable from collagen balls. RESULTS: The structures displayed 3-dimensionality, a hyalinized core and a covering of benign-appearing cells,fratures identical to those seen in collagen balls from serous cavity fluids. Averaging 94 x 60 microm, they were similar in size to a human oocyte at ovulation. Their presence did not correlate with any concurrent Pap smear findings or previous medical history. CONCLUSION:Since collagen balls have previously been reported to occur only in serous cavity fluids, they are thought to arise at those locations. We propose that collagen balls are transported via the fallopian tube and uterine fundus to the cervix, much in the same way that an ovum is transported to the uterine flindus for implantation. The likelihood of such an event is most probably quite low. The recent finding of 7 cases of collagen balls may correlate with an increased prevalence of collagen balls in peritonealfl uids, thus increasing the probability of their transport to the cervix.     

Cytology of peritoneal washings in gynecologic patients. Diagnostic criteria and pitfalls

A total of 226 peritoneal washing specimens obtained from gynecologic patients over a three-year period was reviewed. The diagnostic problems encountered were: differentiating reactive mesothelium from low-grade malignancies; distinguishing between benign ovarian tumors, ovarian tumors of borderline malignancy and low-grade ovarian malignancies; and potential false-positive diagnoses in endometriosis. Low-grade malignancies could be distinguished from reactive mesothelium by evaluating subtle cytologic criteria and by comparing the cells to those of the tumor on histologic section. The differentiation of low-grade epithelial malignancies from benign epithelial lesions caused difficulties that could only be resolved by evaluating the histologic material. Positive peritoneal washings increased the stage in 8 of 110 patients undergoing initial surgery for gynecologic malignancy. Two of 76 patients undergoing a second-look laparotomy had positive washings without histologic evidence of tumor. These ten patients did less well than did those with similar histology but negative washing cytology, despite receiving additional therapy because of the cytologic findings.     

Peritoneal cytology in uterine papillary serous carcinoma.

OBJECTIVE: To highlight the significance of positive peritoneal cytology in uterine papillary serous carcinoma (UPSC). STUDY DESIGN: Seventeen consecutive UPSC cases with peritoneal cytology from 1993 to 1997 were reviewed and compared with the original cytologic diagnosis and extent of tumor involvement in tissues. RESULTS: Of the 17 post-menopausal women with UPSC, 11 had early-stage tumors (clinical stage I and II); three cases (27%) with positive peritoneal cytology were upgraded from at least International Federation of Gynecologists and Obstetricians stage IA to IIIA. No change in surgical stage was noted in four of six (67%) advanced cases with positive peritoneal cytology. The review diagnoses of peritoneal cytology did not differ from the original diagnoses. CONCLUSION: The features of UPSC in peritoneal cytology are those of a high grade malignancy and may be shared by tumors with similar histology from other sites. The malignant features are readily identified, but the site of origin may not be completely ensured. Positive peritoneal cytology upgrades the surgical stage of early-stage UPSC cases and helps with prognostication and treatment. One case with positive washings but without residual tumor probably represented early spread and/or multicentric origin of the tumor.     

Multiple myeloma. The diagnostic role and prognostic significance of exfoliative cytology

The clinical significance and diverse cytomorphologic spectrum of exfoliative cytology in multiple myeloma are presented from our 20-year retrospective and continuing prospective studies and from an extensive review of the literature. Of 370 myeloma patients studied retrospectively, 126 had at least one exfoliative cytologic specimen but only 6 had one or more specimens positive for myeloma. These included six pleural and two ascitic fluids and one sputum. In Papanicolaou-stained smears, myeoloma cells varied from essentially normal-appearing plasma cells to dispersed large malignant cells with little or no plasmacytoid features. Whereas all 203 cervical or vaginal, cerebrospinal, urine and bronchial specimens were negative for myeloma, 40% and 50% of the pleural and ascitic fluids, respectively, were positive. Four prospectively studied patients produced a total of seven positive serous fluid specimens. Follow-up data was available for eight patients with cytology positive for myeoloma. Six were dead within three months of the first positive specimen.     

Müllerian inclusions in peritoneal washings. Potential source of error in cytologic diagnosis

Müllerian inclusions in peritoneal washings from female patients may be mistaken for adenocarcinoma. Such findings were studied in the peritoneal washing cytology specimens from eight cases. The inclusions usually presented as tubular or papillary structures, often forming a single layer of epithelium surrounding psammoma bodies. The cells forming these structures often displayed some degree of atypia. Recognition of this entity in peritoneal fluids is important to avoid a misdiagnosis of disseminated cancer. A general outline is proposed for interpreting such findings in peritoneal washings, based on the cytomorphology of these structures as well as the microscopic features of the primary neoplasms.     

Prospective comparative cytologic study of direct peritoneal smears and lavage fluids in patients with epithelial ovarian cancer and benign gynecologic disease

Direct peritoneal samples obtained by scraping or brushing (with a Cytobrush) were compared to peritoneal lavages (washings) for the cytologic evaluation of patients with gynecologic disease. The direct samples were obtained during laparotomy or laparoscopy, following saline lavage if that was performed, and were immediately smeared on glass slides and fixed in 95% alcohol. Only 9 of the direct peritoneal samples taken from 64 patients with benign gynecologic disease were unsatisfactory for cytologic interpretation while 19 of the 33 lavage specimens simultaneously collected from these patients were considered unsuitable for analysis (P less than .001). Two direct smears from cases with benign histology were reported as suspicious. Nineteen patients with epithelial ovarian cancer also had cytologic specimens collected by direct sampling and by washing. The direct smears were positive for malignancy in 12 cases, suspicious in 4 cases and negative in 3 cases while the lavage samples were positive in 9 cases, suspicious in 4 cases, negative in 4 cases and unsatisfactory in 2 cases. These results indicate that direct peritoneal sampling is a simple and reliable alternative to peritoneal lavage and produces a significantly lower incidence of unsatisfactory specimens.     

Peritoneal washings in endometrial carcinoma. A study of 298 patients with histopathologic correlation

OBJECTIVE: To correlate findings of peritoneal washings in patients with endometrial carcinoma with histologic parameters. STUDY DESIGN: Between 1995 and 1998, 298 women with endometrial carcinoma were treated by hysterectomy with intraoperative peritoneal washings (PW) at Memorial Sloan-Kettering Cancer Center. All cytology and pathology slides were available for review. Pathologic parameters of hysterectomy specimens were evaluated and correlated with the findings of PW. RESULTS: Thirty-two patients (10.7%) had abnormal PW. Two hundred sixty-two had endometrioid adenocarcinoma; 26 of them had abnormal PW (10.0%). Thirty-six patients had other histologic subtypes (papillary serous carcinoma, clear cell carcinoma and adenosquamous carcinoma), and six of them had abnormal PW (16.7%). The incidence of abnormal PW in the two groups was not significantly different (P = .78). Among 26 patients with endometrioid adenocarcinoma and abnormal PW, there were 17 cases (9.9%) of International Federation of Gynecology and Obstetrics (FIGO) grade 1, 7 (12.7%) of grade 2 and 2 (5.7%) of grade 3 (P = .56). Ten cases (14.9%) had no myometrial invasion, 10 (7.0%) had myometrial invasion of < or = 50% of myometrial thickness, and 6 (11.5%) had invasion of > 50% of myometrial thickness (P = .18). Vascular invasion was present in 8 cases (14.8%) and absent from 17 (8.2%) (P = .14). Eighteen patients (7.6%) had stage I/II disease, and eight patients (30.8%) had stage III/IV disease (P = .001). Among 298 patients, cervicovaginal smears performed before surgery were available for review in 76. Five of the 7 patients (71.4%) with abnormal PW and 37 of the 69 patients (53.6%) with normal PW had abnormal Pap smears (P = .45). CONCLUSION: Abnormal PW did not correlate with histologic subtypes, FIGO grade, depth of myometrial invasion, vascular invasion or abnormal Pap smears. A significantly higher incidence of abnormal PW was associated with stage III/IV disease.     

P-3THE VALUE OF PERITONEAL WASHING CYTOLOGY IN THE STAGING OF GYNAECOLOGICAL MALIGNANCY

Introduction:  Current protocols for staging gynaecological cancers include cytopathological examination of peritoneal washings taken at the time of definitive surgery. We investigated the clinical usefulness of this procedure.
Methods:  During 2004 and 2005, 140 peritoneal washings were submitted for cytopathological examination in our institutions for staging of 36 ovarian, 101 endometrial and 3 synchronous ovarian/endometrial cancers.
Results:  The washings contained malignant cells in 39 cases (28%). 35 of these cases had high stage disease – not confined to the organ of origin (i.e. stage 2 or more for ovary and stage 3 or more for endometrial). The other 4 were stage 1C ovarian cancers where there was either rupture or tumour involvement of the capsule. In only 2 of the 39 positive cases the cancer was marginally upstaged by the positive washings – these were ovarian cancers upstaged from 2A /B to 2C.
Discussion:  These findings suggest that peritoneal washing cytology as a routine procedure for staging ovarian and endometrial cancer is of limited clinical value. A larger study is needed to determine whether this procedure should continue to be included in staging protocols for gynaecological cancer.     

Bethesda 2001. Impact on the reporting of gynecologic cytology

OBJECTIVE: To examine the impact of implementing Bethesda 2001 in one laboratory. STUDY DESIGN: A computer search identified all cervicovaginal specimens evaluated between July 2001 and June 2002. Bethesda 2001 was implemented on January 1, 2002. The rates of specimen adequacy and the frequency of each diagnostic category 6 months before and 6 months after the implementation of Bethesda 2001 were compared. RESULTS: A total of 21,332 cervicovaginal specimens were evaluated during the study period. During the first 6 months, 10,695 specimens were examined; 40% were liquid-based preparations. During the next 6 months, 10,367 specimens were examined; 60% were liquid-based preparations. Prior to the implementation of Bethesda 2001, the percentages of each category were as follows: 74.99% within normal limits, 7.10% reactive/reparative cellular changes (R/R), 10.29% atypical squamous cells (ASC), 0.24% atypical glandular cells (AGC), 3.45% low grade squamous intraepithelial lesion (LSIL), 3.44% high grade squamous intraepithelial lesion (HSIL) and 0.73% unsatisfactory. In addition, 19.00% were classified as "satisfactory but limited by" (SBLB). Following the implementation of Bethesda 2001, the percentages of each category were as follows: 80.09% negative for intraepithelial lesion and malignancy including 6.94% with the qualifier R/R, 10.32% ASC, 0.27% AGC, 4.54% LSIL, 3.44% HSIL and 0.81% unsatisfactory. In addition, 17.40% were satisfactory with a quality indicator (SAT with QI). The incidence of reporting benign endometrial cells in patients over age 40 was the same for both periods. There was a significant decrease in the percentage of specimens classified as SAT with QI when compared to that of specimens classified as SBLB. A statistically significant increase in the percentage of specimens was noted in the category LSIL (P < or = .001) and satisfactory (.005) after implementing Bethesda 2001. No significant changes were noted in other categories. CONCLUSION: Our laboratory experienced some changes in the laboratory statistics of reporting gynecologic cytology after the implementation of Bethesda 2001. Continuous monitoring of reporting trends is indicated to clearly understand the impact of Bethesda 2001 on laboratory statistics.     

Liquid-based cytology improves productivity in cervical cytology screening.

Objectives: The ThinPrep test was introduced into our institution on a phased basis over 3 years between January 2002 and December 2004. This study set out to assess its effect on productivity (as measured by output of cases per medical scientist per day) during the changeover period. Numbers of high and low-grade lesions and of unsatisfactory slides were also monitored. Methods: The percentage conversion from conventional preparation to liquid-based cytology (LBC) and output of cases per medical scientist per day were calculated from our database at 6-month intervals. The average backlog, average number of cases received per month and percentage of unsatisfactory and abnormal cases were calculated similarly. Results: Over the study period 92 084 cases were received. The percentage of cases using ThinPrep increased: from 9% in January 2002 to 73% in December 2004. During the study there was an increase in output from 17.0 to 22.3 cases per medical scientist per day, representing a 31% improvement at 73% conversion. Numbers of unsatisfactory cases decreased substantially and the numbers of low and high-grade diagnoses were relatively constant. Conclusions: The change to ThinPrep has improved productivity and decreased the number of unsatisfactory cases. There was no adverse effect on quality during the changeover.     

The sensitivity of detection of asbestos bodies in sputa and bronchial washings

While asbestos bodies (ABs) in sputum and/or bronchial washings are highly specific markers for significant asbestos exposure, comparison of the sensitivity between sputum cytology and bronchial washing cytology for the detection of ABs had not been documented. Review of the files of the Cytopathology Laboratory, Methodist Hospital, Houston, Texas, for the period 1973 to 1984 identified 11 patients with slides available for review who (1) had been examined by both sputum cytology and bronchial washing cytology and (2) had at least one specimen positive for ABs. Of the 11 evaluable cases, all had ABs in the bronchial washings but ony 6 had ABs in the sputum. In addition, iron stain (e.g., the Prussian blue stain) was found to be more sensitive than the Papanicolaou stain for the detection of ABs in these cases. These findings indicate that iron-stained bronchial washing specimens should be preferred for the cytologic detection of asbestos exposure.     

Factors significant in the diagnostic accuracy of lung cytology in bronchial washing and sputum samples. I. Bronchial washings

Some factors influencing the diagnostic accuracy for primary lung cancer in bronchial washings were studied in 276 consecutive cases seen between 1959 and 1974. Diagnostic accuracy increased during the years under study; the reasons included increasing expertise of the laboratory staff, better documentation of cytologic criteria and improved collection techniques. The overall accuracy was 74%. Detection of malignant cells was highest for squamous-cell and adenosquamous carcinomas (81%), small-cell carcinoma, adenocarcinoma and large-cell carcinoma (70%) and lowest for bronchioloalveolar-cell carcinoma (47%). Accuracy was 84% for central tumors as compared to 30% for peripheral lesions. Tumors of less than 2 cm in diameter yielded very poor results (15%) while those greater than 2 cm yielded 82% accuracy. The specificity of diagnosis of cell type in those specimens with malignant cells was over 93% for squamous-cell carcinoma, small-cell carcinoma and adenocarcinoma, 77% for large-cell carcinoma and below 50% for adenosquamous carcinoma, bronchioloalveolar carcinoma and the uncommon tumors. Two bronchial washings per case gave an appreciably better result (92%) than one per case (68%). The percentage of unsatisfactory specimens from those with cancer was 13.5 and from a control group was 29.9. Reasons for unsatisfactory specimens included limited cellular material, excessive blood and/or leukocytes and drying artifacts.