Rapid stereological method for assessing the area fraction (A_A) of prostatic carcinoma in TURP specimens

Document(s) associé(s)

Annexes

• AS_17_n2_283_ZAI (2,6M)

Abstract

Incidental carcinoma of the prostate (Stage A) is reported in 10% of TURP (Transurethral resection) specimens. The aims of this study were first, to assess the accuracy of the area fraction (A_A) of prostatic cancer defined as the percentage of chips against stereology (gold standard technique) and second to evaluate a newly described rapid stereology technique based on drawing tube and image analysis system. Fifty formalin fixed fully embedded TURP specimens from patients with prostatic carcinoma were used. Patients were divided into three groups in relation to the A_A of carcinoma. Group 1 (20) the A_A was 0-5%, group 2 (18) the A_A was 5.1-35% and group 3 (12) the A_A was 35.1-100%. H&E stained slides were used to assess the A_A by three different techniques. Regression line analysis and the Altman and Bland correlation were used for the comparison between the methods. The overall accuracy of A_A (0-100%) of both routine and new methods was good in comparison with stereology. In group 1 the percentage of chips method was less sensitive than stereology (p=ns). The new method compared well with stereology (P<0.0001) technique in all three groups of patients. The Altman and Bland correlation showed overestimation of the percentage of chips which was more marked at the level of upper value for group 3 with overlap between groups 1 and 2. The new method showed minor underestimation for the lower values for group 1 and small overestimation for the upper value in group 3. In general the new method compared well with stereology technique in patients with carcinoma of prostate. It is concluded from this study that routine counting of chips with tumour was less sensitive than stereology for the accurate assessment of A_A. The new drawing tube method compared well with stereology, and it was fast and accurate for assessing A_A of tumour in TURP specimens in patients with incidental (stage A) and advanced (stage B) carcinoma. The Altman and Bland correlation was superior to the coefficient of variation for assessing the differences and agreement between two methods.

Pour citer cet article

A propos de : Abd Al-Mouti Zaitoun