Premium
A new strategy based on fluorodeoxyglucose‐positron emission tomography for managing liver metastasis from colorectal cancer
Author(s) -
Watanabe Akira,
Harimoto Norifumi,
Araki Kenichiro,
Yoshizumi Tomoharu,
Arima Kota,
Yamashita Yoichi,
Baba Hideo,
Tetsuya Higuchi,
Kuwano Hiroyuki,
Shirabe Ken
Publication year - 2018
Publication title -
journal of surgical oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.201
H-Index - 111
eISSN - 1096-9098
pISSN - 0022-4790
DOI - 10.1002/jso.25250
Subject(s) - medicine , standardized uptake value , positron emission tomography , perioperative , colorectal cancer , metastasis , primary tumor , radiology , pathological , fluorodeoxyglucose , cancer , nuclear medicine
Background Prognostic models are needed to manage liver metastasis from colorectal cancer (CRLM). Thus, we developed an algorithm to guide treatment based on the standardized uptake value (SUV) from fluorodeoxyglucose‐positron emission tomography (FDG‐PET). Methods We retrospectively evaluated 148 patients who underwent surgery for CRLM, including 107 cases of primary surgery and 41 cases with preoperative chemotherapy before conversion surgery. We evaluated the prognostic value of perioperative SUV among primary surgery cases, as well as the prognostic value of the SUV change ratio after conversion surgery (postchemotherapy/prechemotherapy SUV). Results In the primary surgery group, recurrence‐free survival (RFS) was independently predicted by an SUV of ≥6.04 ( P = 0.042) and ≥4 liver metastases ( P = 0.003). The combination of an SUV of ≥6.04 and ≥4 liver metastases was strongly associated with poor RFS ( P < 0.001). In the conversion surgery group, the SUV change ratio was associated with tumor size change and pathological response. An SUV change ratio of ≥0.293 was associated with shorter RFS ( P = 0.006) and independently predicted RFS ( P = 0.026). We established a therapeutic algorithm for managing CRLM based on these results. Conclusion FDG‐PET may be useful for predicting recurrence and prognosis in cases of CRLM, and our algorithm may be useful for managing multiple CRLMs.