Abstract
348
Introduction: Colorectal Cancer (CRC) follow-up involves carcinoembryonic antigen (CEA) measurements, and combined imaging such as 18-FDG PET/CT scans. Considering cost vs. patient benefit, many protocols (ASCO, NCCN) recommend imaging in cases with serially elevated CEA. Since CEA is neither highly sensitive nor specific, FDG-PET scans are often critical.
Objectives: Our goal was to: 1) Assess if a correlation exists between sites of FDG-avid CRC recurrences and rising CEA titers/ doubling times. 2) Comment on factors that may play a role in this relationship. Methods: Patients (74% M, 26% F, mean age: 62Y) of biopsy proven CRC on follow-up were retrospectively identified. Those having FDG- PET/CT scans and serial CEA measurements were included in the study, with CEA doubling times calculated in months with a standard formula. FDG-avid metastases were taken as recurrences. 79% of the patients were post-surgical with 21% receiving only chemotherapy. Tumors were staged (TNM), and diagnosis was divided by location (55% rectal, 23% Colon, 14% sigmoid, 7% others). Recurrences were divided into Hepatic and Extrahepatic (local, skeletal, pulmonary, intracranial). Details such as surgery type, and chemotherapy regimen were collected to assess differences in response of CEA doubling time and recurrence on PET/CT. The data was tested for normality and analysis was done using appropriate statistical tests.
Results: 36% of the sample population (n=70) did not have serially rising CEA titers which reinforces the borderline sensitivity of the test. As a trend, greater stages of CRC corresponded to a larger initial CEA values (Kruskal Wallis Test: P value- 0.184) but there was no correlation found between Tumor Stage and CEA Doubling time (P-0.384). All stages of CRC had similar percentages of rising CEA titers (T2: 66%, T3: 66%, T4: 63%), and upon further nodal and metastatic breakdown, no significant correlation existed between higher TNM and proclivity towards rising CEA (independent sample Chi square test: P- 0.466). Significant correlation existed between final diagnosis and rising CEA titers. Rectosigmoid CA (100%) and Sigmoid CA (80%) were most likely to have rising CEA levels while Rectal CA had rising titers only 51% of the time (Chi square test: P-0.022). CEA doubling times differed depending on the chemotherapy regimens used, with Capecitabine (25m) having the highest average time and slowest disease progression, while those receiving no chemotherapy (-21m) had a faster disease progression (Kruskal Wallis: P: 0.419). No significant correlation was found between sites of FDG avid metastasis on the PET/CT and CEA doubling times, with hepatic and extra-hepatic metastasis sites compared separately (Mann- Whitney test: P- 0.76, 0.927)
Conclusions: While FDG PET successfully identified CRC recurrences without serial CEA elevations, there was no significant link between site of FDG avid metastatic recurrence and CEA doubling times/ rising titers. The significant correlation between final diagnosis and rising CEA titers may be useful while planning imaging. Sigmoid and rectosigmoid carcinomas are more than 80% likely to have rising CEA titers while solely Rectal carcinoma was only 51% as likely, indicating a bigger role for PET/CT in the latter cases.
CRC Subtype Diagnosis vs. CEA titers