Prognostication in cancer of unknown primary (CUP): Development of a prognostic algorithm in 311 cases and review of the literature

详细信息    查看全文

• 作者：Dimitrios Petrakis ; George Pentheroudakis ; Evangelos Voulgaris ; Nicholas Pavlidis
• 关键词：Cancer of unknown primary ; Prognostic factors ; Algorithm ; Survival
• 刊名：Cancer Treatment Reviews
• 出版年：2013
• 出版时间：November, 2013
• 年：2013
• 卷：39
• 期：7
• 页码：701-708
• 全文大小：674 K

文摘

| Figures/TablesFigures/Tables | ReferencesReferences

Summary

Background

CUP represents a heterogeneous population of patients with systemic malignancy and variable outcomes. Identification of clinical, pathologic and laboratory parameters with prognostic utility could contribute to estimation of death hazard and tailoring of therapy.

Patients and Methods

Clinical, pathologic and laboratory data from 311 patients with CUP diagnosed in a single university centre from 1988 to 2011 were examined for prognostic significance in univariate, multivariate and Classification and Regression Tree (CART) analyses. We analysed all published CUP prognostic algorithms in PubMed and EmBase from 1985 to date in order to describe defining characteristics.

Results

Most patients harboured poorly differentiated adenocarcinoma or carcinoma (85 % ) in visceral sites (62 % ) and were managed with combination chemotherapy. Median overall survival for all patients was 8 months (95 % CI 6.7-9.1). Multivariate analysis established that only leucocytosis (HR 0.37, p = 0.001, cut off <10,000/mm³ leucocytes), clinicopathologic CUP subgroup (HR 2.44, p = 0.001 for the visceral subgroup) and performance status (HR 0.58, p = 0.002 for PS 0-1) retained independent prognostic significance. These three parameters were used for developing a prognostic algorithm (Ioannina Score for CUP Outpatient Oncologic Prognostication, I-SCOOP) which produced a dynamic 5-tier point score and classified patients in low, intermediate and high risk groups with median survival times of 36, 11-14 and 5-8 months respectively. We identified 15 published CUP series describing prognostic algorithms with common, as well as distinct, patient characteristics and prognosticators.

Conclusions

We developed a simple and easy to use CUP prognostic algorithm based on readily available clinicopathologic and laboratory variables. However, analysis of all published series revealed lack of prognosticator consensus, highlighting the heterogeneity of the disease.