**4. Discussion**

The present study demonstrated that the operative extent of patients with unilateral multifocal PTC was not associated with risk of recurrence. Various guidelines and recent

publications have promoted a "less is more" approach for the treatment of low-risk PTC, which represents the vast majority of thyroid cancers; this involves less extensive operation, less radioiodine, and less or no thyroid hormone suppression [23]. A global trend toward less radical surgical procedures, including thyroid lobectomy, has also gained traction in recent years [24]. Thyroid lobectomy has several advantages, including lowering the risk of complications and possibly avoiding a lifelong need for thyroid hormone supplements; however, concerns about oncological safety remain for patients with specific risk factors, including multifocality [25].

Determining the optimal surgical extent for patients with multifocal PTC has been a long-standing problem [26–28]. A consensus report of the European Society of Endocrine Surgeons recommended total or near-total thyroidectomy for multifocal PTC patients to reduce local recurrence [12]. A meta-analysis further indicated that patients with multifocal PTC should undergo central LN dissection [29]. On the contrary, recent studies demonstrated comparable RFS between total thyroidectomy and lobectomy for patients with node-negative multifocal PTC [18,19]. This inconsistency may be partly attributed to multifocality-associated risk factors, including large tumor size and LN metastasis, which can affect the operative extent decision [7]. These factors should also be considered to determine the optimal surgical extent for multifocal PTC.

Total thyroidectomy is usually performed instead of lobectomy for patients with more aggressive clinicopathological characteristics [30]. In the present study, compared with the thyroid lobectomy group, the total thyroidectomy group was also found to have a larger tumor size, a higher rate of microscopic ETE, and an increased risk of LN metastasis. As these high-risk features could affect the development of recurrence, propensity score matching was performed to minimize potential biases [21]. After adjusting for possible confounding factors, including tumor size, microscopic ETE, LN metastasis, and coexisting Hashimoto thyroiditis, our matched cohorts showed no difference in RFS between the total thyroidectomy and lobectomy groups. Our results further confirmed the overall oncologic safety of thyroid lobectomy for patients with multifocal PTC.

Subgroup analyses were performed to determine whether thyroid lobectomy is feasible for multifocal PTC patients with various risk factors. Total thyroidectomy showed comparable RFS to that of thyroid lobectomy for patients with non-PTMC (*p* = 0.711), node-positive PTC (*p* = 0.536), and ATA intermediate risk of recurrence (*p* = 0.682). Our findings suggest that multifocal PTC patients with tumor size > 1 cm, LN involvement, or intermediate ATA risk of recurrence do not always require a more extensive operation. Multivariable Cox proportional hazards analysis also indicated that the operative extent was not associated with the risk of recurrence, regardless of other risk factors. Therefore, we believe that thyroid lobectomy is suitable for all multifocal PTC patients without high-risk factors that require total thyroidectomy.

Our study has some limitations. First, this study was a retrospective cohort study, which is prone to selection bias. Patient selection for total thyroidectomy might be influenced by various factors, and the assignment of thyroid lobectomy and total thyroidectomy was not randomized. Although we performed propensity score matching, the results might be influenced by selection bias. Second, we did not evaluate long-term outcomes such as cancer-specific survival. During the mean follow-up period of 5.2 years, there was no cancer-specific mortality in the present study. Third, the number of patients with other risk factors of recurrence was relatively small. When we calculated the sample size for survival analysis using the data from Table 3, more than 255 recurrences were required to determine the statistical significance with HR of 1.686, 90% of power, and 2.5% of significance. Further validation studies with a larger cohort and long-term follow-up are warranted.
