Frozen sections--a retrospective study.

Two hundred and forty-four (244) specimens were examined by frozen section (FS) from 1987 through 1988 at the Kidwal Memorial Institute of Oncology. Of these, eleven (11) were performed for evaluation of the surgical margin of tumours and fifty four (54) for detection of lymph node metastasis. All these examinations (65) proved to be hundred percent (100%) accurate. The remaining 179 frozen sections were performed for the diagnosis of an unknown pathologic process. Of these, 44.69 percent were precisely diagnosed, in 41.9 percent the pathologic process was correctly but not precisely diagnosed, in 11.17 per cent the diagnosis was deferred and remaining 2.24 percent were incorrectly diagnosed with no harmful consequences to the patients. By eliminating the cases where diagnosis was deferred and combining results of "precise diagnosis" with those of "correct pathologic process", the overall accuracy rate was 98.2 percent. Thus, FS has greater benefit when used for the general diagnosis of an unknown pathologic process rather than for an exact or precise diagnosis.