3. Results
3.1. Baseline characteristics of the patient sample
Between February 1, 2009, and February 1, 2017, 130 patients were included, with 99 patients in the very low-risk group and 31 patients in the low-risk group, according to the CHIC-HS criteria. Totally, there were 26 and 104 patients included in the US and NC groups, respectively. All patients received adjuvant chemotherapy postoperatively.
The baseline demographic and clinical characteristics before and after PSM are shown in Table 2. Before PSM, there were no significant differences in sex and serum AFP levels between the NC and US groups (p =0.21 and p =0.437, respectively), but SMD was 0.264 for sex and 0.189 for serum AFP levels. The mean age of the patients was higher in the US group than in the NC group, although the difference was not significant (702±533 vs. 561±525, mean±standard deviation,p =0.098, SMD=0.269). In contrast, the distribution of PRETEXT stages differed significantly between the groups (p =0.001, SMD=0.849). After PSM, all SMD values were <0.2. In the NC and US groups, the mean age was 528.0±324.8 days and 603.7±522.1 days, respectively (p =0.851), with 50% male participants in each group (p =1). The distribution of PRETEXT stage and serum AFP levels was similar between the groups (p =0.655 and p =0.806, respectively). Overall, patient characteristics were similar for all variables included in PS matching.
3.2. Association between NACT, surgical outcomes, and pathological findings
Postoperative pathological results are shown in Table 3. A positive surgical margin was observed in three (13.6%) and six (27.3%) patients in the US and NC groups, respectively. There was no significant difference between the groups (p =0.268). There were significant differences in pathological classification between the groups (p =0.031). The pathological classification in the NC and US groups was pure fetal histology (PFH) in five (22.7%) and two (9.1%) patients, epithelial mixed embryonal/fetal (EMEF) in nine (40.9%) and 16 (72.7%) patients, mixed epithelial-mesenchymal (MEM) in eight (36.4%) and two (9.1%) patients, and small cell undifferentiated (SCU) in zero (0%) and two (9.1%) patients, respectively. The number of EMEF cases was significantly higher in the NC group than in the US group.
3.3. Patient outcomes
The median follow-up period was 64 months (95% confidence interval [CI]: 58.2–65.5): US group, 62.18 months (interquartile range [IQR]: 60–69) and NC group, 68.82 months (IQR: 60–75.75). In the US group, the 5-year EFS and OS were 81.8% (95% CI: 60.9–93.3) and 86.3% (95% CI: 65.8–96), respectively. Events occurred in four (16.7%) patients, all of whom experienced an event <14 months after surgery (Table S1). One patient with EMEF pathology died during chemotherapy from sepsis, two SCU patients and one patient with EMEF pathology experienced recurrence, and the lesion in one patient with SCU recurred as a hepatic sarcoma. Three patients underwent a second surgery after relapse, and the tumors were resected. The patient with EMEF pathology survived, and the two SCU patients died due to progression of lung metastases.
In the NC group, the 5-year EFS and OS were 81.8% (95% CI: 60.9–93.3) and 90.9% (95% CI: 71.0–98.7), respectively. Events occurred in four (16.7%) patients, all of whom experienced an event <9 months after surgery (Table S1). One patient with EMEF pathology died due to multiple organ failure, and three patients with MEM pathology developed recurrence and underwent a second surgery; one patient died of hepatic metastases, and two patients had long-term survival after surgery. No significant difference was observed between the groups in terms of EFS and OS (p =0.964 andp =0.655, respectively) (Fig. 1). Our analysis of the putative risk factors of interest showed that resection margin status, PRETEXT stages, age, sex, and AFP levels were not associated with death or events, but that pathological classification was significantly associated (p =0.007 and p=0.032, respectively). However, no events occurred in patients with PFH pathology, and they survived until the last follow-up, whereas all SCU patients died, and the sample size was thus underpowered to calculate odds ratios and conduct further analyses.