High rates of histopathologic discordance in sarcoma with implications for clinical care.

Chandrajit P. Raut, S. George, Jason L. Hornick, James E. Butrynski, J. A. Morgan, John E. Ready, Alessandra F. Nascimento, Chris Fletcher, G. D. Demetri, E Baldini

Journal of Clinical Oncology · 2011 · 15 citations · 0 references

Concepts

Abstract

10065 Background: Histology-specific therapy for sarcoma patients (pts) at dedicated sarcoma centers is routinely guided by re-review of histopathology (HP) from referring institutions. We evaluated the rate of discordance between HP diagnoses rendered by the referring institution and our institution and the impact on treatment. Methods: This retrospective study included all pts with a new diagnosis of primary/metastatic sarcoma with available HP slides and reports referred to our sarcoma center January-December 2010. Rates of discordance and impact on treatment and prognosis were assessed. Univariate analyses were performed to evaluate the association of discordance with HP characteristics. Results: Of 795 unique pts referred to our sarcoma center, 335 (42%) met inclusion criteria; ineligible pts included 195 (25%) with recurrence, 92 (12%) without a biopsy, and 36 (5%) from our institution. HP discordance was noted in 79 cases (24%), of which 52 (16%) were deemed clinically significant, altering treatment recommendations (systemic therapy regimen, extent of surgery, use of radiation) and/or prognosis. Using the referring institution’s HP diagnosis as the reference, low discordance rates were noted for GIST (2/49, 4%), angiosarcoma (0/25, 0%), and Ewing sarcoma (1/9, 11%). In contrast, discordance rates were higher for liposarcoma (all subtypes, 5/23, 22%), leiomyosarcoma (7/34, 21%), and rhabdomyosarcoma (5/6, 83%). Ten cases (13%) were diagnosed as “sarcomas” at referring institutions and as non-sarcoma lesions (e.g., lymphoma, melanoma, T11 fracture callus) at our institution. HP materials on 102 pts (30%) were sent for second opinion review prior to referral to our center; 2 cases were discordant. On univariate analysis, referring institution diagnoses of fibroblastic tumors (p=0.04) or tumors with skeletal muscle differentiation (p<0.0001) were significantly more likely to be discordant. Conclusions: Rates of HP discordance in sarcoma can be high and may impact therapeutic decisions in a substantial subset of pts. This may in part reflect the bias of challenging cases referred to a sarcoma center. Nonetheless, pts with the more frequently discordant sarcoma subtypes may benefit from second opinion review.