Efficacy of NACT
The study was based on data from 774 patients with cervical cancer enrolled from 28 hospitals in China between January 2016 and October 2019. Consent was withdrawn by 25 patients and 12 patients did not meet the inclusion criteria. Therefore, 369 and 368 patients were randomly assigned to the NACT and PST groups, respectively. 320 patients in the NACT group and 328 patients in PST group were included in per-protocol analysis (Fig. 1). The number of patients lost to follow-up was 21 (5.9%) in the NACT group and 16 (4.3%) in the PST group.
To give a clearer view of the efficiency of NACT, we compared the baseline characteristics of the NACT and PST groups. Although the patients were randomly assigned to the two groups, the baseline demographics and clinical characteristics were not well balanced between the two groups in terms FIGO stage and tumor size (Additional file 1: Table S1). The NACT group presented a higher rate (62.6%) of patients with carcinoma invasion beyond the uterus than the PST group (44.6%, p < 0.005). Tumor size was also larger in the NACT group than in the PST group (p = 0.002), indicating that patients in the NACT group had more advanced tumors. 71.4% (252/303) patients achieved CR/PR, including 30 patients achieving pCR.
By February 2022, 73 months had passed since the first patient was enrolled, and the median (interquartile range [IQR]) follow-up time was 34 (26–43) months. In the Kaplan–Meier analysis, patients in the NACT group showed a similar rate of PFS to those in the PST group (HR 0.828; 95% CI, 0.574–1.254; p = 0.320; 5-year PFS, 80.5% vs. 83.8%;). Similarly, OS in the NACT group was no worse than in the PST group (HR 0.750; 95% CI, 0.426–1.321; P = 0.319; 5-year OS, 85.8% vs. 87.8%) (Fig. 2).
Characteristics and Survival Outcomes of Patients Undergoing NACT
We further performed a subgroup analysis of the NACT group based on NACT sensitivity; 227 (70.9%) patients were in CR/PR group, and 93 (29.1%) patients were in SD/PD group. No significant differences between the two groups were found with respect to the baseline characteristics accessed before receiving NACT (Additional file 1: Table S2). The median age of the patients was 48 years (range: 20–64 years), and the median BMI was 22.8 kg/m2 (range: 15.8–34.5 kg/m2). Additionally, there were no significant differences between laboratory examination results between the two groups including serum hemoglobin concentration and platelet concentration. The distribution of FIGO stage and pathological type was well balanced. All responders and 91 of 93 non-responders received RS after NACT, and all these patients had completed tumor pathology. The pathological classification was identified as follows: squamous cell carcinoma (288 cases, 90.6%), adenocarcinoma (22 cases, 6.9%), and squamous adenocarcinoma (eight cases, 2.5%). There were no significant differences between the CR/PR and SD/PD groups with respect to histological subtype—deep cervical invasion (33.9% and 44.0%, respectively), lymphovascular space invasion (12.3% and 15.4%, respectively), parametrial involvement (0.9% and 3.3%, respectively), vagina involvement (1.8% and 3.3%, respectively), uterus involvement (4.8% and 7.7%, respectively), and lymph node metastasis (10.6% and 15.4%, respectively). Only one case in the CR/PR group had a positive vaginal resection margin (Additional file 1: Table S3).
For all patients undergoing NACT, the 5-year OS rate was 86.5% (86.9% for responders and 85.5% for non-responders), and the 5-year PFS rate was 81.4% (82.1% for responders and 79.9% for non-responders). By February 2022, 44 of 320 (13.8%) patients in the NACT group experienced disease progression or death from any cause. Among them, 25 of 227 (11.0%) patients in the PR/CR group, and 13 of 93 (14.0%) patients in the SD/PD group developed recurrences. The PR/CR group had fewer distant metastases than the SD/PD group (four cases, 1.8% vs. three cases, 3.2%, respectively). Moreover, a total of 25 deaths had been reported, with 16 (7.0%) patient deaths in the PR/CR group, and nine (9.7%) patient deaths in the SD/PD group (Additional file 1: Table S4). No significant difference was found in PFS and OS between responders and non-responders (PFS: HR, 1.197 [95% CI, 0.635–2.258], p = 0.579; OS: HR, 1.459 [95% CI, 0.644–3.302], p = 0.365, Fig. 3A and 3B). In addition, two deaths unrelated to cancer had been reported. Specifically, one patient died of COVID-19 and one committed suicide. These two patients belonged to the PR/CR group. Therefore, we plotted the survival curves of disease-specific survival (DSS) and found that non-responders did not have poorer survival outcomes (DSS: HR, 1.699 [95% CI, 0.722–3.857], p = 0.231, Fig. 3C).
Prognostic Significance of Clinicopathological Factors for the SD/PD group
To further analyze the clinicopathological risk factors affecting the prognosis of patients with locally advanced cervical cancer in the NACT group and its subgroups (CR/PR and SD/PD groups), we performed univariate analysis. As shown (Additional file 1: Table S5), for all patients undergoing NACT, lower hemoglobin concentration (Crude OR, 0.980 [95% CI, 0.966–0.994], p = 0.005), larger tumor size before NACT (Crude OR, 1.253 [95% CI, 1.011–1.552], p = 0.039) or after NACT (Crude OR, 1.255 [95% CI, 1.064–1.481], p = 0.007), > 1/2 depth of cervical invasion (Crude OR, 3.359 [95% CI, 1.808–6.243], p < 0.001), and lymph node metastasis (Crude OR, 2.350 [95% CI, 1.160–4.763], p = 0.018) were associated with worse PFS. For the CR/PR group, lower hemoglobin concentration (Crude OR, 0.977 [95% CI, 0.961–0.994], p = 0. 0.007), lower platelet count (Crude OR, 1.004 [95% CI, 1.001–1.008], p = 0. 0.021), larger tumor size before NACT (Crude OR, 1.372 [95% CI, 1.042–1.808], p = 0.024) or after NACT (Crude OR, 1.880 [95% CI, 1.333–2.651], p < 0.001), > 1/2 depth of cervical invasion (Crude OR, 3.375 [95% CI, 1.624–7.017], p = 0.001), uterus involvement (Crude OR, 3.448 [95% CI, 1.201–9.897], p = 0.021), and lymph node metastasis (Crude OR, 3.982 [95% CI, 1.822–8.700], p = 0.001) were associated with worse PFS. For the SD/PD group, the clinicopathological risk factor for PFS was > 1/2 depth of cervical invasion (Crude OR, 3.303 [95% CI, 1.011–10.791], p = 0.048). When the candidate variables with a p-value of < 0.2 in univariate analysis of Cox proportional hazards regression were included in the multivariable model, deep cervical invasion was the only significant independent prognostic factor related to poor PFS for all patients undergoing NACT (Adjust. OR, 2.580 [95% CI, 1.310–5.079], p = 0.006). Additionally, deep cervical invasion was associated with poorer PFS in responders (Adjust. OR, 2.358 [95% CI, 1.092–5.090], p = 0.029). In addition, patients with lower hemoglobin concentrations (Adjust, OR, 0.976 [95% CI, 0.9570.996], p = 0.020), larger tumor size after NACT (Adjust. OR 1.720 [95% CI, 1.085–2.726], p = 0.021), and lymph node metastasis (Adjust. OR, 2.391 [95% CI, 1.0385.508], p = 0.041) had significantly worse PFS in responders. Nevertheless, when regarding non-responders to NACT, no clinical or pathological factors were associated with PFS (Table 1).
Table 1
Multivariate Analysis of Progression-free Survival for Patients Undergoing NACT
| CR/PR (n = 227) | | SD/PD (n = 91) | | NACT (n = 318) |
| Adjust p-Value | Adjust OR (95%CI) | | Adjust p-Value | Adjust OR (95%CI) | | Adjust p-Value | Adjust OR (95%CI) |
Age | | | | 0.308 | 0.967 (0.906, 1.032) | | | |
Hb | 0.020 | 0.976 (0.957, 0.996) | | \ | \ | | 0.121 | 0.987 (0.972, 1.003) |
Plt | 0.459 | 1.002 (0.997, 1.006) | | \ | \ | | 0.357 | 1.002 (0.998, 1.005) |
BMI | \ | \ | | 0.201 | 0.892 (0.749, 1.063) | | 0.568 | 0.971 (0.877, 1.075) |
Tumor size | | | | | | | | |
before NACT | 0.566 | 0.904 (0.642, 1.274) | | \ | \ | | 0.923 | 0.987 (0.760, 1.282) |
after NACT | 0.021 | 1.720 (1.085, 2.726) | | \ | \ | | 0.281 | 1.119 (0.912, 1.372) |
> 1/2 depth of cervical invasion | 0.029 | 2.358 (1.092, 5.090) | | 0.108 | 2.751 (0.803, 9.180) | | 0.006 | 2.580 (1.310, 5.079) |
LVSI | | | | \ | \ | | 0.857 | 0.927 (0.409, 2.102) |
Vagina involvement | \ | \ | | \ | \ | | 0.679 | 1.393 (0.290, 6.690) |
Uterus involvement | 0.233 | 1.998 (0.641, 6.225) | | \ | \ | | 0.627 | 1.306 (0.445, 3.838) |
Lymph node metastasis | 0.041 | 2.391 (1.038, 5.508) | | \ | \ | | 0.310 | 1.478 (0.696, 3.140) |
Candidate variables with a p-value < 0.2 on univariate analysis of Cox proportional hazards regression were included in the multivariable model and “\” meant the corresponding factor was not included. |
Abbreviations: OR, odds ratio; Hb, hemoglobin concentration; Plt, platelet concentration; BMI, body mass index; SCCA, Squamous Cell Carcinoma Antigen; NACT, neoadjuvant chemotherapy; LVSI, lymphovascular space invasions |
Next, we performed a univariate analysis affecting OS (Additional file 1: Table S6). For all patients undergoing NACT, the risk factors of OS were consistent with PFS, including lower hemoglobin concentrations (Crude OR, 0.979 [95% CI, 0.961–0.998], p = 0.029), larger tumor size before NACT (Crude OR, 1.456 [95% CI, 1.131–1.874], p = 0.004) or after NACT (Crude OR, 1.386 [95% CI, 1.140–1.685], p = 0.001), > 1/2 depth of cervical invasion (Crude OR, 3.832 [95% CI, 1.683–8.962], p = 0.002), and vagina involvement (Crude OR, 5.567 [95% CI, 1.303–23.790], p = 0.021). For responders, lower hemoglobin concentrations (Crude OR, 0.976 [95% CI, 0.953–0.999], p = 0.039), larger tumor size before NACT (Crude OR, 1.509 [95% CI, 1.080–2.106], p = 0.016) or after NACT (Crude OR, 1.945 [95% CI, 1.298–2.914], p = 0.001), > 1/2 depth of cervical invasion (Crude OR, 3.575 [95% CI, 1.299–9.840], p = 0.014), and lymph node metastasis (Crude OR, 3.785 [95% CI, 1.313–10.906], p = 0.014) were associated with worse OS. No factor was found resulting in worse OS for non-responders. However, when we analyzed the correlation between OS and factors by multivariate analysis, besides deep cervical invasion was associated with worse OS for all patients undergoing NACT (Adjust. OR, 2.608 [95% CI, 1.017–6.691], p = 0.046), no other clinicopathological factors were found associated with OS in multivariate regression, whether it was for responders, non-responders, or all patients undergoing NACT (Table 2).
Table 2
Multivariate Analysis of Overall Survival for Patients Undergoing NACT
| CR/PR (n = 227) | | SD/PD (n = 91) | | NACT (n = 318) |
| Adjust p-Value | Adjust OR (95%CI) | | Adjust p-Value | Adjust OR (95%CI) | | Adjust p-Value | Adjust OR (95%CI) |
Age | \ | \ | | 0.108 | 0.925 (0.841, 1.017) | | \ | \ |
Hb | 0.093 | 0.978 (0.953, 1.004) | | \ | \ | | 0.300 | 0.989 (0.968, 1.010) |
Tumor size | | | | | | | | |
before NACT | 0.760 | 0.926 (0.563, 1.522) | | 0.945 | 1.029 (0.457, 2.319) | | 0.650 | 1.082 (0.771, 1.518) |
after NACT | 0.055 | 1.928 (0.987, 3.769) | | 0.463 | 1.306 (0.641, 2.660) | | 0.096 | 1.246 (0.962, 1.614) |
Approach of surgery | | | | | | | 0.085 | 0.305 (0.079, 1.176) |
FIGO | 0.272 | 1.920 (0.599, 6.152) | | \ | \ | | \ | \ |
> 1/2 depth of cervical invasion | 0.154 | 2.243 (0.739, 6.806) | | 0.325 | 2.357 (0.428, 12.97) | | 0.046 | 2.608 (1.017, 6.691) |
LVSI | \ | \ | | \ | \ | | 0.242 | 1.901 (0.648, 5.577) |
Vagina involvement | 0.523 | 0.427 (0.031, 5.831) | | 0.165 | 6.154 (0.472, 80.171) | | 0.676 | 1.476 (0.238, 9.157) |
Uterus involvement | \ | \ | | \ | \ | | 0.927 | 1.076 (0.227, 5.109) |
Lymph node metastasis | 0.464 | 1.591 (0.46, 5.507) | | \ | \ | | 0.938 | 1.042 (0.368, 2.946) |
Candidate variables with a p-value of < 0.2 in univariate analysis of Cox proportional hazards regression were included in the multivariable model and “\” meant the corresponding factor was not included. |
Abbreviations: OR, odds ratio; Hb, hemoglobin concentration; Plt, platelet concentration; BMI, body mass index; SCCA, Squamous Cell Carcinoma Antigen; NACT, neoadjuvant chemotherapy; LVSI, lymphovascular space invasion. |
Overview of post-NACT therapy for non-responders
In the 28 different centers, the treatment of responders was performed according to the experimental design, while the treatment of the non-responders varied according to the experience of attending physicians. Previous analyses have shown that patients who were not sensitive to NACT also achieved survival outcomes that were not inferior to those of NACT-sensitive patients, and we consider that the prognosis here may be related to a different choice of treatment regimen. Therefore, we focused our analysis on the treatment regimens for non-responders and explored whether and how the treatment affected the prognosis of non-responders.
Detailed information on the treatment and survival outcome of 93 patients insensitive to NACT was given in Fig. 4 and Table 3. Of these patients, 91 of 93 (97.8%) non-responders received RS plus pelvic lymphadenectomy following NACT, and only two patients received sequential chemoradiation (SCRT) after NACT. For two SCRT patients, tumor general metastasis to the cervix, segment VIII of the liver, and the middle and lower lobes of left lung were identified in one patient (PFS = 15), and she died 23 months after entering the experiment (P276 in Table 3); another patient survived disease-free throughout the follow-up period.
Table 3
Characteristics and Treatment Details of Patients Insensitive to NACT who Died or Developed Recurrent Disease
Patient No. | Age | FIGO stage | Histology | Deep cervical invasion | LVSI | Lymph node status | Risk level | Therapy after surgery | Patterns of disease recurrence | Therapy after recurrence | Time to recurrence (months) | Time to death (months) |
P366 | 47 | IB2 | SCC | Positive | Negative | Negative | Intermediate | Chemoradiation | Pelvic Cavity | Untreated | 4 | 6 |
P294 | 47 | IB2 | SCC | Positive | Positive | Negative | Intermediate | None | Pelvic Cavity | Untreated | 5 | 8 |
P133 | 44 | IIA2 | SCC | Negative | Negative | Negative | Low | Chemotherapy | Unknown | Unknown | Unknown | 12 |
P137 | 42 | IIA2 | SCC | Negative | Negative | Negative | Low | Chemotherapy | Sigmoid Colon | Radiotherapy | 9 | 16 |
P080 | 44 | IIA2 | SCC | Positive | Negative | Negative | Intermediate | Immu+ Chemoradiotherapy | Cervical Stump | Chemoradiation | 3 | 24 |
P143 | 51 | IB2 | SCC | Positive | Negative | Negative | Intermediate | Chemotherapy | Unknown | Unknown | 7 | 25 |
P219 | 32 | IB2 | SCC | Positive | Positive | Negative | Intermediate | Chemoradiation | Uterine Cervix | Radiotherapy | 29 | 29 |
P053 | 34 | IIA2 | SCC | Positive | Negative | Positive | High | Chemoradiation | Retroperitoneal Lymph Node | Chemoradiation | 19 | 45 |
P241 | 47 | IB2 | SCC | Positive | Negative | Negative | Low | Chemotherapy | Bladder | Radiotherapy | 7 | 35, Alive |
P167 | 43 | IIA2 | SCC | Negative | Negative | Negative | Low | Chemotherapy | Cervical Stump | Reoperation + Immu. | 19 | 52, Alive |
P121 | 39 | IIA2 | SCC | Positive | Positive | Negative | Intermediate | Chemotherapy | Cervical Stump | Radiotherapy | 20 | 39, Alive |
P373 | 54 | IB2 | SCC | Positive | Negative | Negative | Low | Chemotherapy | Pelvic Cavity | Radiotherapy | 24 | 59, Alive |
P292 | 46 | IIA2 | SCC | Negative | Negative | Negative | Low | Chemotherapy | para-aortic lymph nodes | NA | 36 | 36, Alive |
P276 | 46 | IIA2 | SCC | NA | NA | NA | NA | NA | Cervix, liver, and lung | Unknown | 15 | 23 |
Patient No.276 received concurrent chemoradiotherapy after neoadjuvant chemotherapy, and had no surgically pathological results. For Patient 292, recurrence was found in the last follow-up (2021-11), so the treatment options and long-term prognosis after recurrence were not available at the time of writing. |
Abbreviations: FIGO, International Federation of Gynecology and Obstetrics; LVSI, lymph vascular space invasion; SCC, squamous cell carcinoma; AC, adenocarcinoma; NA, not available; Immu., immunotherapy. |
Detailed operation information on patients undergoing NACT is shown (Additional file 1: Table S7). Since the result of the LACC Trial (26) was not available at the time of the implementation of this clinical trial, the majority of patients underwent a minimally invasive hysterectomy (228 cases, 74.5%). In SD/PD group, the open surgery rates were 31.8%, which was slightly higher than that in CR/PR group (23.1%; Chi-square test p = 0.143). It may result from the difficulty of SD/PD patients’ surgery. During the operation, intraoperative hemorrhage (p = 0.302) and length of parametrium excision (p = 0.233) for the two groups showed no significant differences. Operative complications were recorded as well and a total of 23 patients (7.2%) suffered from at least one kind of operative complication including injury to the urinary system, injury to the gastrointestinal tract, vascular injury, infection, deep vein thrombosis, and lymphocyst with a diameter more than five centimeters. The rate of any operative complications at the time of the analysis was 7.0% in the CR/PR group and 7.7% in the SD/PD group. After surgery, the median indwelling catheter time in the CR/PR group (five days) was shorter than that in SD/PD group (six days; Chi-square test, p = 0.033), and the median amount of post-operation drainage in the CR/PR group (390 mL) was less than that in the SD/PD group (550 mL; Chi-square test, p = 0.060).
In 91 patients receiving radical hysterectomy following NACT, three patients did not receive any postoperative adjuvant therapy, two of whom survived throughout the follow-up period. Patient No.294 refused any treatment after surgery and even after recurrence and died 8 months after being enrolled in the clinical trial. The vast majority of patients (80.2%) received adjuvant chemotherapy, and their chemotherapy regimen was consistent with NACT. Among them, eight patients relapsed (P121, P133, P137, P143, P167, P241, P292, P373 in Table 3) and three patients died. A total of 15 patients underwent radiotherapy with (2 patients) or without (13 patients) platinum-containing chemotherapy or immunotherapy as adjuvant therapy. Among them, recurrence was observed in four patients and they all died from disease-specific causes (P366, P080, P219, P053). Of the five patients that survived recurrence, three received radiotherapy after recurrence, one had a reoperation, and one patients’ treatment regimen was not available.
Prognostic significance of therapeutic factors for SD/PD group
To better distinguish the effect of different adjuvant therapies on the prognosis of the 91 NACT-insensitive patients who received NACT followed by RS, we performed Kaplan–Meier survival analysis and log-rank test and found that only the chemotherapy group had a significantly better PFS and OS than the radiotherapy group (PFS, p = 0.049; OS, p = 0.001) (Fig. 5A and 5B). Considering that different pathological results would affect the choice of treatment, patients were divided into two groups based on risk according to the Sedlis criteria: (1) low-risk and (2) intermediate and high risk group using recognized high-risk pathologic features after radical hysterectomy (27).
In the low-risk group (n = 57), two patients did not receive any adjuvant therapy, two patients received only radiotherapy, and three patients received sequential chemoradiation. None of the above seven patients had cancer progression or died. The remaining 50 patients received only chemotherapy as adjuvant therapy, and among them six patients had died or had disease progression (Table 3). Based on this, we found that adjuvant treatment was not a significant independent prognostic factor for PFS (p = 0.644; no adjuvant treatment vs. chemotherapy alone, p = 0.574; chemotherapy alone vs. radiotherapy, p = 0.453) or OS (p = 0.884; no adjuvant treatment vs. chemotherapy alone, p = 0.774; chemotherapy alone vs. radiotherapy, p = 0.685) in the low-risk group (Fig. 5C and 5D), indicating that additional postoperative treatment did not help to improve the prognosis of patients in the low-risk group.
Among the 34 patients in the intermediate or high-risk group, 23 patients received only chemotherapy, seven patients received sequential chemoradiation, two patients received concurrent chemoradiation, one patient received immunotherapy plus chemoradiotherapy, and one patient refused any adjuvant therapy. Detailed information of death or recurrence is given in Table 3. Postoperative treatment appeared to be beneficial for patients in the intermediate- and high-risk group. We found that the use of chemotherapy decreased disease progression and increased survival compared with radiotherapy (PFS, p = 0.019; OS, p < 0.002) (Fig. 5E and 5F).