Q & A

April 2004

Q. What steps must be taken to validate a histology lab’s decision to modify a staining method?

A. CLIA ’88 and the CAP Laboratory Accreditation Program require laboratories to have procedures for evaluating the quality of the analytical testing process of each method. For every method, the laboratory must verify accuracy, precision, sensitivity, specificity, and reportable ranges, as applicable. Not all of these factors apply to histochemical stains, but the same general principles of quality control and test validation do apply—before implementing a change to an existing procedure, the laboratory must test the proposed changes to ensure the results are reliable.

For most laboratory tests, measuring accuracy requires testing control samples with known amounts of the test substance. For histochemical stains, this means taking steps to verify that modified stains identify the same cellular or histochemical components as the original stain. To assess test accuracy, the lab should stain appropriate tissue control slides to determine whether the modified stain identifies the target—for example, bacteria, mucin, or collagen. Ideally, these control slides should be from tissues shown by previous testing to contain the specified target.

Precision is mainly a measure of test variability and is usually determined through replication tests. To determine whether within-run and between-run variability are acceptable, the lab can stain a series of slides in parallel with the previous method, preferably on at least two occasions. It is also preferable for at least two observers to independently review and document the results. The number of parallel slides required to be stained is arbitrary and must be determined by the laboratory director.

Complete documentation of the steps taken to validate the new method is also required. It is recommended that labs create a validation protocol, which is a written plan stating how such testing will be conducted. This protocol should include a record of the results obtained and the laboratorians evaluating those results, and it should be maintained with the stained study slides. The validation study should also be referenced in the modified technical procedure.

Sue E. Lewis, BS, HTL(ASCP)
University of Iowa Hospitals and Clinics
Iowa City
Member, CAP
Histotechnology Committee

Patrick L. Fitzgibbons, MD
St. Jude Medical Center
Fullerton, Calif.

Advisor, CAP
Histotechnology Committee

Q. Is the Miller disk of great value? Should I purchase it for manual reticulocyte counts?

A. A critical aspect of manual reticulocyte counts is assessing a sufficient number of red cells to ensure that the measurement is adequately precise. The minimum sample size for erythrocytes, within which reticulocytes must be enumerated, is 1,000, as defined by Poisson statistics (CAP hematology checklist item HEM.35300). Manually counting this number of cells obviously is labor intensive.

The Miller disk is a counting aid that provides a standardized area in which to count erythrocytes. The disk incorporates a large square counting area within which is a square that is one-ninth the size. The number of erythrocytes in the large square is calculated by counting all erythrocytes, including reticulocytes, in the small square and multiplying by nine. This is followed by counting all reticulocytes in the large square. The latter number is then divided by the calculated number of erythrocytes in the large square and multiplied by 100 to derive the reticulocyte percentage. To ensure adequate analytic precision, therefore, at least 112 erythrocytes must be present in the small square (equivalent to 1,008 erythrocytes in the large square).

While the Miller disk is a labor-saving device that eliminates the need to manually count all of the erythrocytes in the total sample, one must remember to count the reticulocytes in the large square to ensure a total erythrocyte sample of sufficient size. When properly used, the Miller disk can improve the overall precision of manual reticulocyte analysis. Thus, this device is of significant value, and any laboratory devoting a large amount of resources to manual reticulocyte analysis should consider purchasing it.

Finally, it should be mentioned that automated or semiautomated reticulocyte counting is available on many hematology analyzers. These techniques provide far better precision for reticulocyte enumeration than manual methods and should be used when possible.

Steven H. Kroft, MD
Department of Pathology
University of Texas Southwestern Medical Center
Dallas
Vice Chair, CAP Hematology/Clinical Microscopy Resource Committee

Q. One of our urologists insists that we do not mention on the pathology report the diagnosis of low-grade prostate intraepithelial neoplasia, or PIN, for prostate needle core biopsies. The physician claims that low-grade PIN is of no clinical importance. Are we violating the pathology standards of care if we omit this diagnosis? Will we be liable if five or 10 years later the patient develops high-grade PIN or carcinoma?

A. Your urologist has a valid point. Well-known experts writing in standard uropathologic texts have said:

• "The diagnosis of low-grade PIN is limited by problems of reproducibility, and separation from simple hyperplasia is difficult. For this reason, and because it is not a marker for carcinoma, making this diagnosis is not recommended. Its presence is not an indication for rebiopsy."¹
• "The high level of interobserver variability with low-grade PIN limits its clinical utility, and many pathologists do not report this finding except in research studies."²

The concern is not with omitting a diagnosis, but making a diagnosis that is poorly reproducible and does not indicate carcinoma or need for further biopsy. Low-grade PIN is an irreproducible, possibly irrelevant entity to which our clinical colleagues may overreact, with detriment to good patient care.

Reference:

1. Ro JY, Grignon DJ, Amin MA, et al. Atlas of Surgical Pathology of the Male Reproductive Tract. Philadelphia, Pa.: W.B. Saunders Co.; 1997:43.

2. MacLennon GT, Resnick MI, Bostwick DG, eds. Pathology for Urologists. Philadelphia, Pa.: W.B. Saunders Co.; 2002:97.

John A. Maksem, MD
Chief of Pathology
Mercy Medical Center
Des Moines, Iowa