Oncology · Medical Literature
Recurrence risk of differentiated thyroid cancer treated with total or hemithyroidectomy: systematic review and meta-analysis.
Abstract
Differentiated thyroid carcinoma (DTC) is traditionally treated with total thyroidectomy (TT); however, hemithyroidectomy (HT) has been increasingly used. Previous analyses on DTC recurrence rate with HT and TT have yielded inconsistent results; this study aimed to bridge this gap. Searches were conducted in PubMed/MEDLINE and Embase. Furthermore, references of previous systematic reviews were screened to extend the search until May 2025. Specific data on recurrence rate were independently extracted from the included studies by two observers and then pooled using a random-effect model. The primary endpoint was the risk difference (RD) in cancer recurrence rate between HT and TT. The secondary endpoints included recurrence prevalence (RP) for TT and HT and odds ratio (OR) comparing HT with TT. A total of 2,556 papers were initially identified, and 46 retrospective studies were included. The overall quality of the studies was satisfactory. In total, 98,604 patients were treated with TT (n = 75,531) or HT (n = 23,073). The proportion meta-analyses revealed a pooled RP of 5.8% in TT and 8.1% in HT. The HT group had a significantly higher probability of DTC recurrence than the TT group (OR = 1.56, 95% CI = 1.47-1.66, I 2 = 85%). The RD meta-analysis revealed that the recurrence rate was significantly higher in the HT group than in the TT group, by 3% (95% CI = 3-4, I 2 = 87%). The DTC recurrence rate following HT is significantly higher than that observed following TT. Thyroid specialists should consider this finding when tailoring initial treatment strategies.
Clinical relevance
In patients with differentiated thyroid carcinoma (DTC), does hemithyroidectomy (HT) confer a meaningfully different risk of cancer recurrence compared with total thyroidectomy (TT)? __SNAPSHOT__ {"studyDesign":"Systematic review and meta-analysis of retrospective observational studies; random-effects model","setting":"Multinational; searches conducted in PubMed/MEDLINE and Embase plus reference screening of prior systematic reviews; 46 retrospective studies included","sampleSize":"n = 98,604 total (TT: n = 75,531; HT: n = 23,073)","population":"Patients with differentiated thyroid carcinoma (DTC) treated surgically; specific histologic subtypes, tumor sizes, risk strata, and nodal status across included studies not further detailed in the abstract","intervention":"Hemithyroidectomy (HT) — surgical resection of one thyroid lobe","comparator":"Total thyroidectomy (TT) — bilateral complete thyroid resection","primaryOutcome":"Risk difference (RD) in cancer recurrence rate between HT and TT","followUp":"Not reported","fundingSource":"Not reported"}
Key findings
Across 46 retrospective studies including 98,604 patients, proportion meta-analyses yielded a pooled recurrence prevalence (RP) of 5.8% in the TT group and 8.1% in the HT group — an absolute risk difference of approximately 2.3 percentage points (calculated). The abstract states that the HT group had a statistically significantly higher probability of DTC recurrence than the TT group; however, the specific risk difference (RD) point estimate, its confidence interval, and p-value are not reported in the abstract. The odds ratio (OR) comparing HT with TT was reported as a secondary endpoint but its numerical value, confidence interval, and p-value are not provided in the abstract. NNT/NNH cannot be calculated from the available data. Adverse event data (e.g., surgical complication rates, hypoparathyroidism, recurrent laryngeal nerve injury) were not addressed in the abstract. Survival outcomes (overall survival, disease-specific survival) were not reported as endpoints in the abstract.
Practice implications
This large meta-analysis — encompassing nearly 100,000 patients — found a statistically significant higher recurrence prevalence in HT (8.1%) versus TT (5.8%), a finding directionally consistent with the biological rationale that residual thyroid tissue can harbor occult multifocal or contralateral disease. However, several important caveats constrain interpretation. First, only the abstract is available; quantitative effect size (RD, OR, CIs, p-values) cannot be independently verified. Second, all 46 included studies are retrospective, introducing substantial selection bias: patients undergoing HT in contemporary practice are systematically selected as lower-risk (smaller tumors, favorable histology, absence of nodal disease), which may paradoxically underestimate the true recurrence disadvantage of HT or, conversely, inflate it if confounding favors TT in higher-risk cases. Third, recurrence definitions, follow-up duration, surveillance intensity, and radioiodine use likely varied considerably across included studies, all of which can artifactually inflate recurrence detection in TT cohorts under more intensive post-operative thyroglobulin monitoring. Fourth, overall survival and disease-specific survival data were not reported as endpoints, so whether the observed recurrence difference translates into a mortality difference remains unestablished. Fifth, subgroup analyses by tumor size, histologic subtype, risk stratification (ATA low/intermediate/high), and nodal status are not described in the abstract, yet these are precisely the variables that guide contemporary surgical decision-making. The 2015 and 2025 ATA guidelines permit HT for tumors 1–4 cm without high-risk features; whether the recurrence disadvantage observed in this pooled analysis is concentrated in guideline-concordant low-risk patients or driven by inclusion of intermediate-/high-risk patients selected for HT in earlier eras is unknowable from the abstract alone. Patients not adequately represented likely include those with tumors >4 cm, clinically node-positive disease, and non-papillary DTC histologies. Practice should not change on the basis of this abstract alone; the complete manuscript including subgroup analyses, heterogeneity statistics, risk-of-bias assessments, and survival data must be reviewed. __LIMITATIONS__ 1. Exclusive inclusion of retrospective studies introduces high risk of selection bias and confounding by indication — patients selected for HT vs. TT differ systematically in risk profile. 2. Recurrence definitions, surveillance protocols, and follow-up duration were almost certainly heterogeneous across 46 studies from different eras and institutions, compromising pooled recurrence prevalence validity. 3. Differential surveillance intensity between TT and HT patients (serum thyroglobulin monitoring is feasible after TT but not after HT without prior remnant ablation) may artifactually elevate biochemical recurrence detection in TT groups, biasing results toward apparent superiority of TT. 4. Quantitative heterogeneity statistics (I²) are not reported in the abstract, precluding assessment of between-study variability. 5. Radioiodine use post-TT was variably applied across studies and may confound recurrence rates attributed to surgical extent alone. 6. Survival endpoints (OS, DSS) were not reported; the clinical importance of the recurrence difference cannot be gauged without mortality data. 7. Subgroup analyses stratified by tumor size, histology, ATA risk category, and nodal status are not described in the abstract. 8. The absolute recurrence prevalence difference (~2.3 percentage points, calculated) may have limited clinical importance in low-risk DTC where disease-specific mortality is less than 1–2%, especially when weighed against documented benefits of HT (lower rates of hypoparathyroidism, recurrent laryngeal nerve injury, and hypothyroidism). 9. Publication bias and inclusion of predominantly institutional retrospective series may overrepresent TT-centric centers. 10. Funding source is not reported, limiting assessment of potential conflicts of interest. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ This meta-analysis of 46 retrospective studies (n=98,604) found a higher pooled recurrence prevalence with hemithyroidectomy (8.1%) versus total thyroidectomy (5.8%) in differentiated thyroid cancer, and the difference was reported as statistically significant; however, quantitative effect sizes with confidence intervals were not available in the abstract, and survival outcomes were not reported. Given the retrospective design, near-certain selection bias, heterogeneous surveillance practices, and absence of mortality data, this finding must be interpreted cautiously and does not by itself justify abandoning guideline-concordant hemithyroidectomy for appropriately selected low-risk DTC patients pending review of the full manuscript.