PREDICTING OUTCOME IN PAPILLARY THYROID-CARCINOMA - DEVELOPMENT OF A RELIABLE PROGNOSTIC SCORING SYSTEM IN A COHORT OF 1779 PATIENTS SURGICALLY TREATED AT ONE INSTITUTION DURING 1940 THROUGH 1989

Background. Multivariate analyses in papillary thyroid carcinoma (PTC) have shown that age, tumor size, local invasion and distant metastasi s are independent predictive variables. This study attempted to define a reliable prognostic scoring system for predicting PTC mortality rat es with 15 candidate variables that included completeness of primary t umor resection but excluded histologic grade and DNA ploidy. Methods. The study group comprised 1779 patients with PTC (followed up for >26, 000 patient-years), divided by treatment dates into 1940 to 1964 (n = 764) and 1965 to 1989 (n = 1015). Cox model analysis and stepwise vari able selection led to a prognostic model initially derived from the tr aining set (n = 764). The initial prognostic score was thereafter vali dated externally with the later (1965 to 1989) ''test'' data set. Resu lts. The final model included five variables abbreviated by metastasis , age, completeness of resection, invasion, and size (MACIS). The fina l prognostic score was defined as MACIS = 3. 1 (if aged less-than-or-e qual-to 39 years) or 0.08 X age (if aged greater-than-or-equal-to 40 y ears), + 0.3 X tumor size (in centimeters), + 1 (if incompletely resec ted), + 1 (if locally invasive), + 3 (if distant metastases present). Twenty-year cause-specific survival rates for patients with MACIS less than 6, 6 to 6.99, 7 to 7.99, and 8 + were 99%, 89%, 56%, and 24%, re spectively (p < 0.0001). Conclusions. Because the five variables neede d for MACIS scoring are readily available after primary operation, suc h a prognostic system could have widespread applicability in assessmen t of PTC.