Reproducibility of cervicovaginal ThinPrep cellularity assessment.

Criteria for judging the squamous cellularity adequacy of cervicovaginal ThinPreps using cell counts in representative fields have been proposed. We compared the interobserver reproducibility in the designation of ThinPreps as satisfactory (SAT), satisfactory but limited by (SBLB), or unsatisfactory (UNS) for interpretation. Three reviewers independently assessed 145 cervicovaginal ThinPreps accessioned from 1997-1999 at the University of Kentucky cytopathology laboratory. Scant specimens were retrieved by computer search and more cellular specimens were added to make 145. Adjacent fields on a diameter including center at 10x and 40x were counted for sufficiently preserved and unobscured cells. A specimen was designated UNS if four or more fields at 10x had less than 10 squamous cells, and SBLB if four or more fields at 40x had less than five squamous cells. All three observers agreed on the adequacy designation in 76% of cases and the overall kappa was 0.73 (near excellent). Complete interobserver agreement was present in 86% (63/73) of UNS cases, 79% (33/42) of SAT, but only 47% (14/30) of SBLB slides. When SBLB and UNS categories were combined, the discrimination from the SAT category was excellent (kappa = 0.87) with pairwise k-values of 0.94-0.97. In conclusion, adjacent field cell count methods to judge ThinPrep cellularity provide near excellent interobserver reproducibility. While we have designated both UNS and SBLB adequacy categories, criteria could easily be revised to set one threshold to separate SAT and UNS ThinPreps.     

Accuracy and reproducibility of telecytology diagnosis of cervical smears. A tool for quality assurance programs

We randomly selected 50 cervical smears (benign, 14; atypical squamous cells of undetermined significance [ASCUS], 5; low-grade squamous intraepithelial lesion [LSIL], 10; high-grade squamous intraepithelial lesion (HSIL), 12; squamous cell carcinoma, 6; adenocarcinoma, 3) and captured 1,181 digital images (518 MB) at a maximum resolution of 1,600 x 1,200 pixels and transmitted them by e-mail. Diagnosis of glass slides and digital images was done independently in a double-blind manner by 3 pathologists and 3 cytotechnologists, commencing with the diagnosis of digital images followed by diagnosis of glass slides 3 months later. The procedure was repeated after 3 months. Diagnoses were recorded as benign, ASCUS or atypical glandular cells of undetermined significance, LSIL, HSIL, squamous cell carcinoma or adenocarcinoma, and "inadequate for diagnosis." Diagnostic accuracy and interobserver reproducibility were analyzed using an intraclass correlation coefficient (ICC), which revealed good interobserver agreement for the first (0.72) and second (0.64) glass slide diagnoses and the first (0.72) and second (0.60) digital image diagnoses. The kappa values for intraobserver variation between first and second glass slide diagnoses and first and second digital image diagnoses showed moderate to excellent agreement. Digital images are suitable substitutes for glass slides; telecytology can be used as an alternative method for the cytologic diagnosis of cervical smears, particularly in quality assurance programs.     

Clinical significance of unsatisfactory conventional pap smears owing to inadequate squamous cellularity defined by the Bethesda 2001 criterion

To determine the incidence of clinically significant lesions in long-term follow-up after a diagnosis of inadequate squamous cellularity using former and new criteria, we reviewed conventional Papanicolaou (Pap) smears (January-December 1998) for adequacy based on the Bethesda System 2001 criterion. Of 23,302 Pap smears evaluated in our laboratory, 114 (0.489%) were classified as unsatisfactory and 245 (1.051%) as "satisfactory but limited by" based on the 10% rule. Follow-up information for 5 years was obtained for 172 patients without a concurrent cervical epithelial abnormality: 25 (14.5%) had squamous abnormalities (atypical squamous cells, 22; low-grade squamous intraepithelial lesion, 2; and high-grade squamous intraepithelial lesion, 1). With the Bethesda System 2001 criterion, 167 (97.1%) of 172 smears had inadequate squamous cellularity and 5 (2.9%) were adequate. No differences in the incidence of squamous abnormalities detected on follow-up were noted between patients with unsatisfactory Pap smears owing to inadequate squamous cellularity and patients with satisfactory and negative smears. Our findings raise the question whether patients with unsatisfactory Pap smears and a negative history of gynecologic diseases require repeated Pap smears within 2 to 4 months as suggested by the American Society for Colposcopy and Cervical Pathology guideline.     

Telecytology: Intraobserver and interobserver reproducibility in the diagnosis of cervical-vaginal smears

Telecytologic diagnosis of cervical-vaginal smears is potentially useful because it could allow more efficient use of cytopathologist resources and expertise. A pathologist in one location could, in principle, review cytotechnologists' findings using a video display hundreds or thousands of miles away. Currently, bandwidth restrictions limit practical implementation of such a system to review of fields that had been selected for review by the cytotechnologist. The purpose of our investigation was to evaluate how well this type of review correlates with a review in which the entire slide is available for examination by the pathologist. We prospectively selected 100 consecutive cervical-vaginal smears over an 11-day period in August 1999. For each smear, 4 to 12 fields containing abnormal cells from each slide were digitally imaged. Each of 3 pathologists reviewed all digitized images and all glass slides. Diagnoses based on selected digitized images were compared with those based on conventional pathologist review. The kappa statistic, a measure of chance-corrected agreement (reproducibility), was calculated in each setting. Overall, intraobserver and interobserver reproducibility of cervical-vaginal smear diagnoses is fair to excellent. The use of remote digital images for pathologist review did not introduce large (2-step) diagnostic disagreements. The disagreement between a pathologist's glass slide and digital diagnoses is less than that for different pathologists reviewing glass slides, although interobserver differences were even greater in the interpretation of digital images.     

Evolution of Pap testing at a community hospital: a ten year experience

Over a 10-year period the Pap program at our community hospital laboratory progressed from the use of 100% conventional smears to predominantly ThinPrep (TP) and then to nearly 100% SurePath (SP). During this period the annual Pap volume grew from less than 20,000 to nearly 50,000. This retrospective review focuses on improvements in high-grade squamous intraepithelial lesion (HSIL) detection rates and specimen adequacy rates during the 10-year period.The laboratory database was used to identify 310,080 Pap test records between 1995 and 2004 inclusive. Pap type differences in HSIL detection rates and specimen adequacy rates were examined using the Cochran-Mantel-Haenszel test, with year as the stratification variable. The Breslow-Day test was used to evaluate the consistency of differences across the years. The overall results are summarized below in tabular format. Results of Combined 10-Year Data for Total Number of Cases, HSIL Rate, and Unsatisfactory Specimen Rate by Pap Method are given: [table: see text]. HSIL detection rates were significantly higher for liquid-based Pap tests compared with conventional smears (P < 0.0001). In addition, SP was associated with higher HSIL detection rates than TP with an estimated common odds ratio (OR) of 1.37 (95% confidence interval (CI), 1.10 - 1.71; P = 0.005), and higher specimen adequacy compared with TP or conventional Pap methods (P < 0.0001).     

Can technology expedite the cervical cancer screening process? A Hong Kong experience using the AutoPap primary screening system with location-guided screening capability

We studied the usefulness of an automated screening instrument for processing Papanicolaou (Pap) smears to determine whether it could speed human examination by recording the time to screen 1,007 cervical Pap smears with an AutoPap primary screening instrument with location-guided screening (LGS) software and by conventional microscopic examination. We also assessed the accuracy of the methods to determine preparation adequacy, and we compared the diagnosis by each method. The AutoPap with LGS satisfactorily determined the adequacy of Pap smears and identified the marked abnormal cells for human examination. An accurate diagnosis was possible when only the marked cells were examined, and this method reduced the screening time to less than half that required for conventional screening. With low-grade squamous intraepithelial and more severe lesions as a threshold, there were 37 cases in the conventional group and 29 cases in the LGS group. With atypical squamous cells of undetermined significance as the threshold, there were 111 cases in the LGS group and 93 cases in the conventional group. The AutoPap with LGS can significantly speed the examination of Pap smears without lowering the detection rate of clinically important lesions, thus helping alleviate the cytotechnologist shortage.     

A quantitative model using mean and standard deviation for evaluation of interobserver agreement in nuclear atypia scoring of breast carcinomas in a protocol study

A quantitative model was devised for the purpose of evaluating interobserver agreement in the judgment of the nuclear atypia score of cancer cells in protocol studies. In this model, the mean nuclear atypia score (mu) among observers was considered to approximate the 'true' nuclear atypia score of the tumor, and with an identical mu the strength of interobserver agreement was evaluated by the difference between the ideal standard deviation (sigma) and the observed sigma in the slide conference session. This model was applied to the data of a total of 216 invasive ductal carcinomas acquired at eight slide conference sessions held for the standardization of nuclear atypia scoring and monitoring of interobserver agreement level for enrolled cases in a surgical adjuvant protocol on breast cancer. Compared with conventional percentage interobserver agreement and kappa statistics, the present model, named the lambda model, was more sensitive in detecting the improvement of interobserver agreement in nuclear atypia scoring with repetition of slide conference sessions. The lambda model was particularly effective for evaluating the session which included many intermediate cases with a score between 1 and 2 or between 2 and 3. Percentage interobserver agreement and kappa statistics could estimate the agreement irrespective of the number of observers, but the agreement tended to be underestimated when these intermediate cases were in the majority. The present method was considered to be complementary to conventional methods for evaluating interobserver agreement of cancer histology in protocol studies.     

Use of the Thin Prep® Pap Test™ in clinical practice

The Thin Prep® Pap Test™ (Cytyc Corp., Boxborough, MA) received approval by the Food and Drug Administration in May 1996 as an alternative to the traditional conventional smear. The present direct‐to‐vial study assessed the utility of thin‐layer technology for cervicovaginal screening in clinical practice. From May 1997–February 1998 (10 mo), 15,006 cervical smears were processed and evaluated; of these, 5,423 (36.1%) were conventional smears (CS) and 9,583 (63.9%) were Thin Prep slides (TP). Both methods were analyzed to compare specimen adequacy and detection rates of cervical lesions. The TP method reduced the “satisfactory but limited by” rate by 97% and the unsatisfactory rate by 63%. For low‐grade squamous intraepithelial lesions (LSILs), TP slides yielded 3.6% (348/9,583) as compared to 0.98% (53/5,423) for CS, an increase of 267%. The TP method detected a threefold increase in the number of high‐grade squamous intraepithelial lesions (HSILs) of 1.0% (100/9,583), as compared to 0.3% (17/5,425) for the CS group. The atypical squamous cells of undetermined significance/squamous intraepithelial lesion (ASCUS SIL) ratio was reduced by 54% in the TP group. In routine usage in our laboratory, the Thin Prep® Pap Test™ yielded a significant increase in the detection of LSILs and HSILs as compared to conventional smears. Specimen adequacy was significantly improved. Diagn. Cytopathol. 1999;20:70–73. © 1999 Wiley‐Liss, Inc.     

Performance of a fluid-based, thin-layer papanicolaou smear method in the clinical setting of an independent laboratory and an outpatient screening population in New England.

BACKGROUND: A patented, fluid-based, thin-layer method for preparation of Papanicolaou (Pap) smears (ThinPrep Pap test) has been reported to be significantly more effective than the conventional smear invented by George Papanicolaou. We tested this position by comparing the cytologic diagnosis and specimen adequacy results obtained using the ThinPrep method with data from conventional Pap smears obtained from a similar population. METHODS: Test results of 56 339 ThinPrep specimens were compared with results from 74 756 conventional smear cases obtained from the same sources in a corresponding period of the previous year. RESULTS: The use of ThinPrep for cervicovaginal cytology produced a 75.14% increase in the detection of low-grade squamous intraepithelial lesions and higher diagnoses. Detection of low-grade squamous intraepithelial lesions increased by 71.65% (from 1.58% to 2.71%), and detection of high-grade squamous intraepithelial lesions increased by 102.54% (from 0.26% to 0.52%). There was a 39.11% decrease in the atypical squamous cells of undetermined significance-intraepithelial lesion ratio (from 2.07 to 1.26). There were also marked decreases in the number of specimens categorized as "satisfactory but limited," owing to obscuring inflammation (-94.34%), obscuring blood (-99.84%), and poor fixation (-99.25%). CONCLUSION: ThinPrep produced increased detection of premalignant precursors while improving specimen adequacy.     

Validation of the use of Pap-stained vaginal smears for diagnosis of bacterial vaginosis

Papanicolaou-stained cervicovaginal smears (Pap smears) are used to screen for cervical cancer. Since there is a lack of consensus in published reports respecting the efficacy of Pap-stained smears in BV diagnostics, there is a need to validate their use for diagnosis of BV. Slides from the international BV00 workshop were Pap stained and independently analyzed by four investigators under a phase-contrast microscope. All workshop slides--whether Pap-stained, Gram-stained or rehydrated air-dried smears--were scored according to the same Nugent classification. The diagnostic accuracy of Pap smears for diagnosis of BV had a sensitivity of 0.85 and a specificity of 0.92, with a positive and negative predictive value of 0.84 and 0.93, respectively. The interobserver weighted kappa index was 0.86 for Pap-stained smears compared to 0.81 for Gram-stained smears, and 0.70 for rehydrated air-dried smears using the mean Nugent score as the criterion standard. Provided that the samples are taken from equivalent locations (the vaginal fornix) and analyzed according to the same scoring criteria, there is no discernable difference in the diagnostic accuracy of the three smear-staining methods. The Pap-stained vaginal smears can be used as a wholly adequate alternative to Gram-stained smears for BV diagnosis.     

Interobserver variation in the reporting of cervical colposcopic biopsy specimens: comparison of grading systems.

AIMS: To assess interobserver variation in reporting cervical colposcopic biopsy specimens and to determine whether a modified Bethesda grading system results in better interobserver agreement than the traditional cervical intraepithelial neoplasia (CIN) grading system. METHODS: One hundred and twenty five consecutive cervical colposcopic biopsy specimens were assessed independently by six histopathologists. Specimens were classified using the traditional CIN grading system as normal, koilocytosis, CIN I, CIN II, or CIN III. The specimens were also classified using a modified Bethesda grading system as either normal, low grade squamous intraepithelial lesion (LSIL) or high grade squamous intraepithelial lesion (HSIL). Participants were also asked to categorise biopsy specimens by the CIN system with the addition of the recently proposed category "basal abnormalities of uncertain significance (BAUS)". The degree of agreement between participants was assessed by kappa statistics. RESULTS: Using the CIN system, interobserver agreement was generally poor: unweighted and weighted kappa values between individual pairs of observers ranging from 0.05 to 0.34 (average 0.20) and from 0.20 to 0.54 (average 0.36), respectively. With the modified Bethesda system, interobserver agreement was better but still poor: unweighted and weighted kappa values ranging from 0.15 to 0.58 (average 0.30) and from 0.21 to 0.61 (average 0.36), respectively. There was little or no agreement between observers in the diagnosis of BAUS. CONCLUSIONS: Interobserver agreement in the reporting of cervical colposcopic biopsy specimens using the CIN grading system is poor. Agreement, while still poor, is better when a modified Bethesda grading system is used. There is little or no consensus in the diagnosis of BAUS.     

Atypical squamous cells of undetermined significance: A cytohistological study in a colposcopy clinic.

Cytohistologic correlation was performed by 3 observers on 100 atypical squamous cells of undetermined significance (ASCUS) cases from a colposcopy clinic. Our objectives were to: 1) subclassify ASCUS cases and determine their clinical significance; 2) assess the independent predictive value of different cytologic parameters for biopsy-proven dysplasia (BPD); and 3) calculate interobserver variability. The prevalence of BPD was 73% in the ASCUS favor dysplasia (AFD) group, and 27% in the ASCUS favor reactive (AFR) group (P 0.001). The sensitivity of cervical cytology (AFD) for detecting BPD was 88.5%. Using multiple logistic regression, only nuclear membrane irregularity was found to be independently predictive of BPD (P 0.0001). The interobserver agreement (kappa) was 0.41. Colposcopic smears were inferior to referring smears in detecting dysplasia, with 67% of patients having a referring diagnosis of dysplasia. In a colposcopy clinic population there is a high prevalence (73%) of BPD in the AFD group. Attention should be paid to nuclear membrane irregularity in the risk stratification of ASCUS cases. Diagn. Cytopathol. 1999;21:211-216.     

Accuracy of reporting endocervical component adequacy--a continuous quality improvement project

Inaccurate reporting of the absence of an endocervical (EC) component on Pap smears often results in slide rescreens, amended reports, clinician dissatisfaction, and sometimes unnecessary repeat smears. Therefore, the accuracy of reporting EC component adequacy was selected as a quality indicator for the laboratory continuous quality improvement program (CQI). The process consisted of problem identification, analysis of the situation, collection of data, implementation of solutions, and evaluation of results. The objective of the study was to determine if the accuracy of reporting EC component adequacy on Pap smears improved after application of such a program. During the first phase, 150 Pap smears originally reported with the absence of an adequate EC component and 150 smears reported with the presence of an adequate EC component were rescreened to measure the baseline accuracy of EC component adequacy reporting. The improvement process was then implemented. A cause-and-effect diagram was developed and root cause was determined. A presentation was then made to the cytology staff. Criteria for EC component adequacy were reviewed, examples were shown, and standardized marking of EC component was implemented. Following improvement actions, a second audit of 150 Pap smears reported with the absence of an adequate EC component as well as 150 smears reported with the presence of an adequate EC component was undertaken to measure change in performance in assessing EC component adequacy. For the baseline rescreening, before initiation of the CQI program, 98% accuracy was achieved with smears that were reported as adequate for EC component present. However, the accuracy with smears reported as absence of an adequate EC component was only 71%, i.e., an adequate EC component was identified in almost 1/3 of these cases on rescreen. After the implementation of improvement actions, the accuracy with smears reported with the presence of EC component remained high (98%) and the accuracy of reporting the absence of EC component was 90%. The difference of the latter before and after the implementation was statistically significant (P = 0.015, z-test). The accuracy of reporting EC component adequacy increased following the CQI process. Using reporting EC component adequacy as an example, we demonstrate that by treating clinical problems as quality control issues and applying basic quality improvement tools, a positive outcome can be effected.     

Assessment of specimen adequacy reproducibility: an Italian experience

Interobserver disagreement on smear adequacy may influence the evaluation of the performance of samplers as well as Pap test sensitivity and follow-up. In 1998, the Italian Group for Cervical Cancer Screening (GISCi) promoted a study on the reproducibility of adequacy criteria using a modified version of the Bethesda system. A set of 200 smears was circulated among six Italian laboratories situated in different parts of the country. For each smear, participants were requested to provide a summary judgment on its adequacy and on the cause(s) of inadequacy, if any. Agreement was measured using kappa-type statistics. The agreement among laboratories was generally good. In comparisons, among five laboratories, kappa values ranged from 0.47 to 0.66. At the consensus meeting on 42 slides, on which at least 2 laboratories dissented from the majority, agreement was reached unanimously for 31 reviewed slides and among 5-6 centers for 11. In this article, some guidance is given in order to attribute to one of the two categories satisfactory/unsatisfactory those smears that have been traditionally considered as candidates for the category of "satisfactory but limited by em leader " (SBLB). New Italian guidelines on adequacy proposed the GISCi are presented and the recommendation is made to eliminate the SBLB category.     

Assessment of static telecytological diagnoses' reproducibility in cervical smears prepared by means of liquid-based cytology

Abstract Objective: This study evaluated the accuracy and reproducibility of telecytological diagnoses proffered on the basis of digitized images from cervical smears prepared by means of liquid-based cytology. Materials and Methods: Representative digital cytological images from a total of 404 cervical smears (benign, 135; atypical squamous cells of undetermined significance, 92; low-grade squamous intraepithelial lesion, 62; high-grade squamous intraepithelial lesion, 87; squamous cell carcinoma, 26; and adenocarcinoma, 2) were uploaded to the CytoTrainer e-learning telecytology platform (developed in the Department of Cytopathology, "ATTIKON" University General Hospital, Athens, Greece) with specific password-protected accounts and were reviewed remotely by four independent board-certified cytopathologists (checking round). Their reports were recorded and classified. After 12 and 24 months, the same representative digital images were transferred in random order to the same cytopathologists and were reviewed again (first and second review rounds, respectively). The cytopathologists' first and second round diagnoses were recorded and compared with their initial ones. Results: Statistical evaluation of cytological diagnoses detected no significant difference in diagnostic accuracy among checking and review diagnoses. The overall interobserver agreement was almost perfect with κ values of 0.79-0.97, whereas intraobserver agreement ranged from almost perfect to perfect with κ values of 0.76-1 in all diagnostic rounds. Conclusions: Digital images of cervical smears can be used for rapid and accurate diagnosis, by diminishing turnaround times and improving small cytology departments' quality indices. Diagnoses made by using static telecytological systems can be as reliable as those made by using conventional microscopy, under the conditions that representative images are taken and that standard cytological diagnostic criteria are applied. Telecytology can be used as an alternative method for the cytologic diagnosis of cervical smears, particularly in quality assurance programs.     

Offline telepathology diagnosis of colorectal polyps: a study of interobserver agreement and comparison with glass slide diagnoses

BACKGROUND/AIMS: Technological advances have produced telepathology systems with high quality colour images and reasonable transmission times. Most applications of telepathology have centred on the remote diagnosis of frozen sections or remote real time expert opinions. This study investigates the reproducibility and accuracy of offline telepathology as a primary diagnostic medium for routine histopathology specimens. METHODS: One hundred colorectal polyps (50 hyperplastic, 50 adenomatous) were presented in a randomised order to five histopathologists as offline images on a telepathology workstation. Six images of each case were used: the slide label, a low power scan of all material on the slide, and four higher magnification views. The times taken to prepare the images, and to make the diagnoses, were recorded. Interobserver agreement was measured with kappa statistics and compared with the glass slide diagnoses. RESULTS: The kappa statistics for the interobserver agreement on the telepathology images lay in the range of 0.90-1.00, which is interpreted as excellent agreement, and were significantly higher than those for the glass slide diagnoses (range, 0.84-0.98; p = 0.001). The median time taken to capture the images for a case was 210 seconds. The median time taken to make a diagnosis from the telepathology images was five seconds, which was significantly shorter than for the glass slide diagnoses (median, 13 seconds; p < 0.0005). CONCLUSIONS: Offline telepathology has the potential to be a primary diagnostic medium for routine histopathology with a high degree of reproducibility and short diagnosis times. Further studies are required to validate offline telepathology for different types of specimens and different operators of the image capture system.     

Reproducibility of the mitosis count in the histologic diagnosis of smooth muscle tumors of the uterus

In order to ascertain the reproducibility of the mitosis count in histologic tumor diagnosis a reference set of 10 microscopic slides from smooth muscle tumors of the uterus was shown to six different pathologists, who were asked to record the number of mitotic figures per 10 high power fields in the most active region of each slide. The results, when tabulated., revealed considerable observer variation in this supposedly objective and quantitative diagnostic criterion. The diagnoses of benignity or malignancy, based on this criterion alone, were unanimous in only four of the 10 cases, whereas the diagnoses based on a constellation of histologic criteria were unanimous in all cases and correlated well with the subsequent clinical evolution of those cases with follow-up data. Possible reasons for the lack of reproducibility of the mitosis count are discussed. We conclude that a qualitative estimation of mitotic activity is helpful as one of the diagnostic and prognostic criteria in the assessment of uterine and other smooth muscle tumors, but that it cannot be used as the sole criterion for distinguishing benign from malignant tumors.     

An abnormal cervicovaginal cytology smear in uterine carcinosarcoma is an adverse prognostic sign: analysis of 25 cases

Carcinosarcoma of the uterus has been poorly characterized on cervicovaginal (Pap) smears, and we examine whether they effectively screen for carcinosarcoma and whether an abnormal Pap smear result has any clinical importance. Twenty-five patients with histologically confirmed carcinosarcoma had a conventional Pap smear shortly before diagnosis. Eleven smears (44%) originally were read as abnormal (malignant or atypical), and 4 additional cases were read as abnormal on retrospective review (15/25 [60%]). All malignant elements were epithelial, and 2 cases (8%) had atypical spindle cells, but no diagnostic sarcoma. Cervical involvement was the only histologic parameter correlating with an abnormal Pap smear result (P = .04). Univariate analysis found stage III or IV disease was an adverse prognostic sign compared with stage I or II disease (mean survival, 8 vs 36 months, respectively; P = .001), and multivariate analysis indicated that an abnormal Pap smear result correlated with worse survival (P = .023). The conventional Pap smear is insensitive (60%) for detecting carcinosarcoma, but when the result is abnormal, the Pap is an important stage-independent adverse prognosticator.     

Semiquantitation of bacteria in sputum gram stains.

In many clinical laboratories, bacteria seen in Gram-stained sputum smears are reported semiquantitatively, using a three- or four-category scale consisting of ratings such as numerous, moderate, rare, and none seen. The consistency with which these categories are assigned was evaluated by repeatedly presenting coded smears to seven experienced microbiology technologists. Technologists rated the same smear twice, pairs of smears prepared from the same specimen, and smears prepared after specimen refrigeration. Agreement was assessed with the weighted kappa test. Semiquantitation of gram-negative rods, gram-positive diplococci, and gram-positive cocci in clusters all showed poor reproducibility (kappa = 0.32, 0.34, and 0.17, respectively). Twenty-four percent of paired ratings differed by two or more categories. Lack of reproducibility was due mainly to the inability of the technologists to render a consistent rating when viewing the same slide on separate occasions (P less than 0.001). Variation in the rating styles of different technologists, differences between smears prepared from the same specimen, and specimen refrigeration tended to further decrease the consistency of ratings, but the reductions were not statistically significant. The quantity of potentially pathogenic bacteria in sputum smears is not estimated consistently with standard microscopy procedures and should not be reported.