Premium
PROGNOSIS FOR SOFT‐TISSUE SARCOMA IN THE LOCOMOTOR SYSTEM
Author(s) -
RYDHOLM ANDERS,
BERG NILS O.,
GULLBERG BO,
PERSSON BJÖRN M.,
THORNGREN KARLGÖRAN
Publication year - 1984
Publication title -
acta pathologica microbiologica scandinavica series a :pathology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.909
H-Index - 88
eISSN - 1600-0463
pISSN - 0108-0164
DOI - 10.1111/j.1699-0463.1984.tb04417.x
Subject(s) - medicine , malignancy , soft tissue sarcoma , prognostic variable , proportional hazards model , grading (engineering) , soft tissue , sarcoma , multivariate analysis , statistical significance , hazard ratio , surgery , clinical significance , retrospective cohort study , primary tumor , population , survival analysis , grading scale , metastasis , cancer , pathology , confidence interval , civil engineering , environmental health , engineering
To identify variables of prognostic importance for soft‐tissue sarcoma in the locomotor system, we performed a retrospective follow‐up study on a consecutive, unselected, population‐based series of 237 patients mainly treated by surgery, 1964–1978. Patients with metastasis at the time of diagnosis were not included. All histologic material was re‐evaluated and histologic malignancy‐grading (four‐grade scale) performed without knowledge of the clinical course. The surgical procedures were classified as marginal and broad excisions. Patient follow‐up ranged between 3‐ and 18 years. Multivariate analysis of the data by Coxs proportional hazard regression techniques disclosed seven negative prognostic variables of primary significance; high malignancy‐grade (IV and III), pain at rest, male sex, increasing age and tumor size, a marginal excision and an extracompartmental tumor site, in order of decreasing relative risk (5.9‐1.9) as regards survival. A secondary variable, that of local recurrence, was then included in the model and was found to have a stronger influence on survival than any of the other variables. Patients with local recurrence had a mortality risk which was 8.3 times that of patients without local recurrence. A risk curve based on the prognostic variables having primary significance was constructed. By this risk curve, patients with very good or very bad prognosis could be identified. The results are important when evaluating the efficiency of different therapies in non‐randomized trials. In such studies the prognostic variables could be used to identify patient groups having comparable prognoses. In addition, patients found to have a good prognosis could be excluded from trials with adjuvant therapy.