ABSTRACT
CONTEXT: There is controversy as to whether papillary thyroid microcarcinoma (PTMC) represents more than one disease entity with different outcomes, requiring different treatment. OBJECTIVES: To compare characteristics, outcomes, and factors associated with prognosis of incidental and nonincidental PTMC. SETTING AND DESIGN: Two reviewers performed searches of online databases (1966-2012), reference lists, and conference abstract books. Longitudinal studies of subjects >16 years old receiving any treatments for papillary thyroid cancer ≤10 mm in size were included. Two reviewers independently screened abstracts and articles, extracted data, and assessed quality of studies using National Institute of Clinical Excellence and PRISMA criteria. RESULTS: Of 1102 abstracts identified, 262 studies were reviewed and 17 studies included, comprising 3523 subjects, with mean follow-up of 70 months and total follow-up of 21 329 person-years. This included 854 subjects with incidental PTMC (follow-up, 4800 person-years; mean tumor size, 4.6 mm [range 3.3-6.7 mm]) and 2669 nonincidental PTMC cases (follow-up, 16 529 person-years; mean tumor size, 6.9 mm [range 5.6-8.0 mm]). The recurrence rate in the incidental group (0.5%; 95% confidence interval [CI], 0-1%, P < .001) was significantly lower than that in the nonincidental group PTMC (7.9%; 95% CI, 5-11%), with an OR of recurrence of 14.7 (95% CI, 5.6-54.8, P < .001) for nonincidental PTMC, compared with incidental PTMC. Lymph nodes were involved in 80% (126/157) of recurrences. On meta-regression, age, sex, size, tumor multifocality, lymph node involvement, and treatment modality were not significantly associated with recurrence. CONCLUSIONS: Our meta-analysis strongly suggests the existence of at least two distinct entities of PTMC. Incidental PTMC has different clinical characteristics and a much lower recurrence rate than nonincidental PTMC, suggesting that management protocols should be re-considered. Additional studies with standardized data collection are required to explore potential differences between subgroups of nonincidental PTMC.
