A total of 1848 patients (median age 65 years, range 27–88 years) were enrolled. In 480 patients (25.9%) with available BMI, median BMI was 28.20 (range: 14.06–55.11). Multidisciplinary team data were available in 776 cases (41%) and the therapeutic approach was discussed in 419 (54%).

Complete datasets were available for age, tumor histology, grade and radiation treatment, and over 75% were complete for toxicity and outcomes.

Table 2 reports details of histology, tumour features, stages and missing data.

After surgery, all patients received adjuvant radiotherapy and 588 (31%) received adjuvant chemotherapy before radiotherapy. Table 3 reports radiotherapy schedules according to risk category.

Notably, IRT alone was prevalent in low risk patients (67%). EBRT followed by IRT were administered to 43% of intermediate risk patients and to 46% of high risk cases.

Local relapses occurred in 106/1761 patients (6%) with available information.

Overall, the 5-year probability of local relapse-free survival was 92.4% (95%CI: 90.8–94). It was 100% (95%CI: 100–100) in the low-risk category, 94.3% (95%CI: 92.5–96.1) in the intermediate risk category and 86.2% (95%CI: 82.5–89.9) in the high risk category (p < 0.0001) (Fig. 1a).

Kaplan–Meier curves for the local relapse-free survival (a), distant metastasis free survival (b), overall survival (OS) (c) and cancer specific survival (CSS) (d) for all patients (left) and for the three risk group levels (right)

Univariate analysis showed risk factors for local relapse were age (HR: 1.042; p < 0.0001), disease stage (HR: 1.150; p < 0.0001), grade (HR: 1.781; p = 0.001) and LVSI (HR: 1.768; p = 0.013) (Table 4).

In multivariate analysis only age (HR: 1.055; p < 0.0001), and disease stage (HR: 1.134; p < 0.0001) were significant (Table 4).

Subdividing by risk classes and adding radiotherapy treatments to multivariate models, an increased risk of local recurrence was detected in intermediate risk patients who received IRT alone (HR: 2.513; 95%CI 1.222–5.169; p = 0.012) compared with the reference treatment (EBRT + IRT), while no differences were detected in high risk patients. The low risk class did not show any event and it was not possible to carry out the analysis.

Distant metastasis occurred in 227/1797 patients (12.6%) with available information.

Overall, the 5-year probability of metastases-free survival was 83.4% (95%CI: 81.2–85.6). It was 94.1% (95%CI: 89.4–98.8), in the low-risk category, 86.4% (95%CI: 83.9–88.9) in the intermediate risk category and 74.1% (95%CI: 69.4–78.8) in the high risk category (p < 0.0001) (Fig. 1b).

In univariate analysis, risk factors for metastasis were age (HR: 1.030; p < 0.0001), disease stage (HR: 1.092; p < 0.0001), grade (HR: 2.048; p < 0.0001) and LVSI (HR: 1.899; p < 0.0001) (Table 4).

In multivariate analysis all variables remained significant (Table 4).

Analyzing the risk classes, in intermediate and high risk groups, EBRT-seq boost was associated with greater risk of metastases than reference treatment (EBRT- IRT) (HR: 2.719; 95%CI 1.143–6.467; p = 0.024 for intermediate risk group and HR: 2.696; 95% CI 1.097–6.624; P = 0.031 for high risk group). In the low risk class only 4 events happened and this made the multivariate model unstable, not allowing the analysis.

Death from any cause was observed in 144/1797 patients (8%) with available information.

Overall, the 5-year probability of OS was 90% (95%CI: 88.2–91.8). It was 94% (95%CI: 87.9–100) in the low-risk category, 93.6% (95%CI: 91.8–95.4) in the intermediate-risk category and 80.7% (95%CI: 76.2–85.2) in the high-risk category (p < 0.0001) (Fig. 1c).

In univariate analysis, risk factors for OS were age (HR: 1.083; p < 0.0001), disease stage (HR: 1.115; p < 0.0001), grade (HR: 2.093; p < 0.0001) and LVSI (HR: 1.517; p = 0.040) (Table 4).

Significance persisted in multivariate analysis for age (HR: 1.103; p < 0.0001), disease stage (HR: 1.131; p < 0.0001) and grade (HR: 1.673; p = 0.006 (Table 4).

When OS in each risk category was correlated with RT type, no significant link emerged.

Cancer-specific deathoccurred in 100/1797 patients (5.6%) with available information.

Overall, the 5-year probability of CSS was 92.5% (95%CI: 90.9–94.1). It was 97.1% (95%CI: 94.9–100) in the low-risk category, 95.7% (95%CI: 94.3–97.1) in the intermediate-risk category and 83.6% (95%CI: 79.2–87.9). In the high-risk category (p < 0.0001) (Fig. 1d).

In univariate analysis, risk factors for CSS were age (HR: 1.072; p < 0.0001), disease stage (HR: 1.158; p < 0.0001), grade (HR: 2.615; p < 0.0001) and LVSI (HR: 1.802; p = 0.014) (Table 4).

In multivariate analysis, age (HR: 1.097; p < 0.0001), disease stage (HR: 1.149; p < 0.0001) and grade (HR: 2.090; p = 0.001) remained significant (Table 4).

When CSS in each risk category was correlated with RT type, no significance emerged.

Data on the incidence of acute and late toxicity related to the type of radiation treatment are shown in Fig. 2.

Acute (left panel) and late (right panel) toxicity incidence according to radiotherapy treatments. 1: EBRT = External Beam Radiotherapy treatment; 2: BOOST EBRT seq = External Beam Radiotherapy treatment with sequential boost; 3: EBRT + SIB = External Beam Radiotherapy treatment with concomitant boost; 4: EBRT + IRT = External Beam Radiotherapy treatment with Interventional radiotherapy boost; 5: IRT escl. = Interventional radiotherapy

All acute toxicity parameters (genitourinary, gastrointestinal, skin and haematological) were highest in patients who received EBRT-SIB and lowest in patients who received only IRT (p < 0.0001; Fig. 3).

Acute toxicity according to radiotherapy treatments. 1: EBRT = External Beam Radiotherapy treatment; 2: BOOST EBRT seq = External Beam Radiotherapy treatment with sequential boost; 3: EBRT + SIB = External Beam Radiotherapy treatment with concomitant boost; 4: EBRT + IRT = External Beam Radiotherapy treatment with Interventional radiotherapy boost; 5: IRT escl. = Interventional radiotherapy

Figure 4a shows late toxicity was highest patients who received EBRT-SIB and lowest in those who were given EBRT-seq Boost (p < 0.0001).

Late toxicity (panel a) and vaginal stenosis (panel b) incidence according to radiotherapy treatments. 1: EBRT = External Beam Radiotherapy treatment; 2: BOOST EBRT seq = External Beam Radiotherapy treatment with sequential boost; 3: EBRT + SIB = External Beam Radiotherapy treatment with concomitant boost; 4: EBRT + IRT = External Beam Radiotherapy treatment with Interventional radiotherapy boost; 5: IRT escl. = Interventional radiotherapy

Figure 4b illustrates vaginal stenosis according to RT schedules showing highest incidence in patients who received EBRT-seq Boost.