Purpose: To assess the prognostic value of the pretreatment potential doubling time (Tpot) in carcinoma of the uterine cervix, relative to other established clinical factors.
Methods and materials: Fifty-two patients with cervical cancer were studied prospectively from March 1991 to October 1993. Pretreatment evaluation included examination under anesthesia and tumor biopsy 6 h following the intravenous administration of bromodeoxyuridine (200 mg). Tpot was determined by deriving the labeling index (LI) and S-phase synthesis time (Ts) using flow cytometry. Six patients were not evaluable and excluded. The remaining 46 patients (average age 55 years) were treated uniformly with radical radiation therapy. There were 39 squamous carcinomas and 7 adenocarcinomas. Fédération Internationale de Gynécologie et d'Obstétrique (FIGO) stages were: Ib and IIa, 12 patients; IIb, 18 patients; III and IV, 16 patients. The median external beam dose was 50 Gy (range, 45-52.8 Gy) delivered in 25 fractions. The median intracavitary dose was 40 Gy (range. 25.5-40 Gy) delivered with a single line source to a point 2 cm lateral of the midline, with a mean dose rate of 0.71 Gy/h. The median overall treatment time was 45 days (range, 34-73 days). As of July 31, 1994, 12 patients had died of disease, and the average follow-up for alive patients was 1.4 years (range, 0.5-3.3 years).
Results: There were 27 tumors with diploid deoxyribonucleic acid (DNA) content and 19 tumors were aneuploid. The median and mean Tpot for the 46 patients were 5.5 and 6.6 days, respectively [range, 2.0-25.6 days; coefficient of variation (CV), 74%]. For 25 patients where Tpot measurements were performed at two separate laboratories, there was a fair correlation (r = 0.74), but systematic differences were detected suggesting that the lack of agreement was not simply due to intratumoral variation. To date, 30 patients remained disease free, while 8 patients had pelvic failure and 9 patients developed distant metastases as the first failure site (1 patient developed both at the same time). In univariate analysis, the only significant prognostic factor for disease-free survival was tumor size (p = 0.004). A short Tpot (or high LI) and long overall treatment time (OTT) were weakly associated with poorer disease-free survival, although not statistically significant (1/Tpot, p = 0.14; LI, p = 0.23; OTT, p = 0.04). Age, FIGO stage, hemoglobin level, S-phase fraction, DNA ploidy, and Ts were not associated with disease-free survival. Multiple regression analysis was not performed because of the relatively small number of patients and short follow-up.
Conclusions: Tpot values determined with current techniques by different laboratories cannot be used interchangeably for the purpose of therapy decisions. Vigorous quality assurance and standardization of the laboratory procedures and analysis methods are important to reduce interlaboratory variation. In this uniformly treated group of patients with cancer of the uterine cervix, traditional clinical prognostic factors remain the most important. Preliminary data suggest that the flow cytometry-determined Tpot and labeling index predict for disease-free survival, although a larger number of patients with longer follow-up is required to assess the true prognostic significance of these assays and to determine if their effect is independent of other clinical factors.