Risk of Hematologic Malignant Neoplasms From Abdominopelvic Computed Tomographic Radiation in Patients Who Underwent Appendectomy

Question  Is there a risk of hematologic malignant neoplasms from abdominopelvic computed tomography (CT) radiation when reverse causation bias is controlled for by defining the reasons for CT scan?

Findings  In this nationwide population-based cohort study that included 825 820 patients who underwent appendectomy for acute appendicitis, there was a substantial excess risk of hematologic malignant neoplasms in the CT-exposed group compared with the CT-unexposed group. The carcinogenic risk was most pronounced in patients aged 0 to 15 years.

Meaning  In this study, perioperative abdominopelvic CT was associated with a higher incidence of hematologic malignant neoplasms.

Importance  Whether computed tomography (CT) radiation is truly carcinogenic remains controversial. Large epidemiological studies that purportedly showed an association between CT radiation and carcinogenesis were limited by confounding by indication and reverse causation, because the reasons for CT examination were unknown.

Objective  To measure the risk of hematologic malignant neoplasms associated with perioperative abdominopelvic CT radiation among patients who underwent appendectomy for acute appendicitis.

Design, Setting, and Participants  This nationwide population-based cohort study used the National Health Insurance Service claims database in South Korea to assess 825 820 patients who underwent appendectomy for appendicitis from January 1, 2005, to December 31, 2015, and had no underlying risk factors for cancer. Patients were divided into CT-exposed (n = 306 727) or CT-unexposed (n = 519 093) groups. The study was terminated on December 31, 2017, and data were analyzed from October 30, 2018, to September 27, 2020.

Exposures  Perioperative abdominopelvic CT examination from 7 days before to 7 days after appendectomy.

Main Outcomes and Measures  The primary outcome was the incidence rate ratio (IRR) of hematologic malignant neoplasms for both groups. The secondary outcomes were IRR of abdominopelvic organ cancers and IRR of all cancers. The lag period was 2 years for the primary outcome and 5 years for secondary outcomes. The IRRs were calculated using Poisson regression models with adjustment for age and sex.

Results  Among the study population of 825 820 patients (52.9% male; median age, 28 [interquartile range, 15-41] years), hematologic malignant neoplasms developed in 323 patients in the CT-exposed group during 1 486 518 person-years and 500 patients in the CT-unexposed group during 3 422 059 person-years. For all hematologic malignant neoplasms, the IRR for the CT-exposed vs CT-unexposed group was 1.26 (95% CI, 1.09-1.45; P = .002). In terms of individual categories of hematologic malignant neoplasms, the CT-exposed group had an elevated risk only for leukemia (IRR, 1.40 [98.75% CI, 1.04-1.87, adjusted by Bonferroni correction]; P = .005). There was no between-group difference in incidence rate of abdominopelvic organ cancers (IRR, 1.07 [95% CI, 1.00-1.15]; P = .06) and that of all cancers (IRR, 1.04 [95% CI, 0.99-1.09]; P = .14).

Conclusions and Relevance  This study controlled for reverse causation bias by defining the reasons for CT scan, and findings suggest that abdominopelvic CT radiation is associated with a higher incidence of hematologic malignant neoplasms. Efforts should be continued for judicious use of CT examinations.