Nomogram based on CT and clinical features to predict R0 resection in patients with stage IIB–IV epithelial ovarian cancer: a multi-center study

Abstract

To develop a nomogram based on CT and clinical features to predict R0 resection in patients with stage IIB–IV epithelial ovarian cancer (EOC). 209 patients with stage IIB–IV EOC from three independent medical institutions were stratified into training cohort (from institutions 1 and 2, n = 144) and independent validation cohort (from institution 3, n = 65). Univariate and multivariate logistic analyses of CT and clinical features obtained within two weeks before debulking surgery were used to determine the independent predictors of R0 resection in the training cohort. Nomogram was developed based on the predictors. Receiver operating characteristic (ROC) curves and calibration curves were performed to evaluate the predictive performance of the nomogram. R0 resection was achieved in 66.00 and 61.50% patients in the training and validation cohorts, respectively. In the training cohort, overall peritoneal cancer index based on CT (CT-PCI) (OR  1.245, P < 0.001), serum human epididymis protein4 (HE4) level (OR   1.003, P = 0.012), and neutrophil-to-lymphocyte ratio (NLR) (OR 1.272, P = 0.031) were independent predictors of R0 resection in patients with stage IIB–IV EOC. Nomogram based on them achieved areas under the ROC curves of 0.908 (95% CI 0.848–0.950) and 0.779 (95% CI 0.659–0.873) in the training and independent validation cohort, respectively. The calibration curves showed good agreement between the nomogram predictions and the actual observations in both cohorts. The nomogram based on overall CT-PCI, serum HE4 level, and NLR could be reliable in predicting R0 resection in EOC patients with stage IIB–IV.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-33657-5.

Introduction

Ovarian cancer (OC) is the third most common gynecologic malignancy worldwide, but it has the highest mortality rate of all gynecologic malignancies¹. As the most common histological type of OC, epithelial ovarian cancer (EOC) accounts for about 90% of OC².

Debulking surgery is the cornerstone of the treatment for patients with EOC. All lesions in the pelvic cavity, abdomen, and retroperitoneum should be resected in patients with pelvic and epigastric involvement (i.e., stage ≥ IIB according to the International Federation of Gynecology and Obstetrics (FIGO) staging system)³. Debulking surgery outcomes are categorized into three types based on residual tumor size: R0 (no macroscopic residual tumor), R1 (residual tumor ≤ 1 cm), and R2 (residual tumor > 1 cm). It has been demonstrated that patients achieving R0 resection have longer progression-free survival (PFS) and overall survival (OS) than those with non-R0 resection^4–6.

For patients with a high probability of R0 resection evaluated before the surgery, primary debulking surgery (PDS) should be performed; otherwise, neoadjuvant chemotherapy followed by interval debulking surgery (IDS) is needed. Hence, identifying patients with a high probability of R0 resection before debulking surgery is significant in clinical practice to ensure individualized and optimal treatment decisions.

Whether a patient achieves R0 resection not only depends on the tumor burden and extent of invasion but also on the patient’s general condition^7,8. CT plays an essential role in evaluating the possibility of achieving R0 resection before both PDS and IDS in OC. Preoperatively, CT helps assess the extent of disease and identify features that may preclude R0 resection, thereby guiding the selection between PDS and neoadjuvant chemotherapy. After neoadjuvant therapy, CT reassessment is crucial to determine whether the tumor has regressed sufficiently for R0 resection and to plan the optimal surgical approach⁹. Some studies have assessed CT’s ability to predict surgical outcomes and developed predictive scoring systems, such as the Suidan score and peritoneal cancer index based on CT (CT-PCI)^10–12. However, the reliability and repeatability of the scoring systems should be improved^13,14. Moreover, except for the Suidan score, the others do not consider the influence of clinical factors on R0 resection^15,16.

Therefore, this study aimed to determine the independent predictors of R0 resection based on CT-PCI, CT features of the primary tumor and metastatic lymph node, and clinical features in patients with stage IIB–IV EOC, and to develop a predictive nomogram model.

Materials and methods

This study was approved by the Medical Ethics Committees of the Affiliated Hospital of North Sichuan Medical College (Institution1, No. 2024ER720-1), the Second School of Clinical Medicine, North Sichuan Medical College (Institution 2, No. 2024 − 176), and the Zigong First People’s Hospital (Institution 3, No. 2024-049). The study was conducted in accordance with the Declaration of Helsinki (2024 version). Due to the retrospective nature of the study, the institutional review boards of the Affiliated Hospital of North Sichuan Medical College, the Second School of Clinical Medicine, North Sichuan Medical College, and the Zigong First People’s Hospital waived the need to obtain informed consent.

Patients

From January 2017 to September 2024, pathologically confirmed EOC patients were consecutively collected from Institution 1, 2, and 3 according to the following inclusion criteria: (1) age > 18 years; (2) initial diagnosis of clinical stage as IIB–IV according to the 2021 FIGO staging system; (3) underwent debulking surgery regardless of PDS or IDS; (4) peripheral hematological parameters, including the serum albumin, total platelet, lymphocyte, monocyte and neutrophil counts, and carbohydrate antigen 125 (CA125), human epididymis protein 4 (HE4) levels, were obtained within two weeks before debulking surgery (either PDS or IDS after NACT); (5) an abdominal CT scan was performed within two weeks before the debulking surgery ( either PDS or IDS after NACT). The exclusion criteria were as follows: (1) incomplete data; (2) combined with other malignant tumors; (3) recent infection; (4) poor CT quality.

209 consecutive stage IIB–IV EOC patients were divided into a training cohort (Institutions 1and 2, n = 144) and an independent validation cohort (Institution 3, n = 65), with cohort allocation detailed in Fig. 1. Clinical, surgical, pathological, and CT data were obtained from institutional databases.

Fig. 1.

Patient collection flowchart.

Clinical data collection

Clinical data (age, FIGO stage, treatment, histology, hematological parameters) were obtained from three institutions. The neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocyte ratio (LMR), and the systemic inflammation response index (SIRI) before the surgery were calculated.

CT scanning protocol

The CT scans were performed with a 64-section or a more than 64-section multidetector CT (Institution1: LightSpeed VCT, GE Medical Systems, USA; or SOMATOM Definition AS+, Siemens Medical Systems, Germany; or UCT710, United Imaging Healthcare, China; Institution2: LightSpeed VCT, GE Medical Systems, USA; Institution3: Revolution CT, GE Medical systems, USA).

All patients underwent non-contrast and contrast-enhanced CT scans within two weeks before debulking surgery, regardless of whether they received IDS after NACT or PDS. Contrast-enhanced scans, including arterial and portal venous phases, were performed 25 and 65 s after a bolus injection of contrast agent (Omnipaque, Iohexol, GE Healthcare, USA; or Ioversol, Hengrui Medicine, China) via an automated pump injector (Vistron CT injection system, Medrad, USA; or Stellant system, Medtronag, Germany) into an antecubital vein. The contrast agent was administered at a total volume of 1.5 mL/kg body weight and an injection rate of 3.0 mL/s, followed by a 20-mL saline flush. The CT scanning parameters were as follows: peak voltage of 120 kV; tube current of 100–250 mA (automatic exposure control); collimation of 64 × 0.6 mm, pitch of 0.9; matrix of 512 × 512 mm, and slice thickness of 5 mm. The coverage of CT scans was from the apex of the diaphragm to the inferior border of the pubis.

Imaging analysis

All CT images before debulking surgery were analyzed independently by two observers (the first and corresponding authors with 2 and 14 years of experience in gynecological imaging, respectively) who were blinded to the surgical outcome and clinical data of the patients.

The following parameters were recorded: (a) primary tumor boundary, clear or unclear; (b) primary tumor location, unilateral ovary or bilateral ovaries; (c) the maximum diameter of the primary tumor; (d) metastatic lymph node (LN) were classified as positive or negative based on the following criteria: short-axis diameter ≥ 1 cm, central necrosis, or fusion/clustering of LNs. (e) CT-PCI was performed according to the criteria proposed by Sugarbaker et al.¹⁷. Based on the Sugarbaker’s PCI system, the peritoneal cavity is divided into 13 separate regions. The first 9 regions (Region 0–8) were divided by two vertical lines through bilateral midclavicular lines, respectively, and two horizontal lines through the anterior superior iliac spines and below the costal margin, respectively. The upper jejunum, lower jejunum, upper ileum, and lower ileum constitute the remaining 4 areas (Region 9–12) (Fig. 2a). In each region, the carcinomatosis is quantified with a score according to the tumor size (Fig. 2b and c). If no lesion is found in the region, a 0 score is obtained. Tumors up to 0.5 cm, up to 5 cm, and > 5 cm or confluence obtain a score of 1, 2, and 3, respectively. The overall CT-PCI is calculated as the sum of lesion scores across the 13 anatomical regions of the peritoneal cavity. Additionally, in order to evaluate the impact of upper abdominal lesions on achieving R0 resection, carcinomatosis in the upper-abdominal regions (region 1–3) was separately analyzed and quantified as the upper-abdominal CT-PCI score. Quantitative data were reassessed by the first author 30 days later to evaluate intraobserver agreement; qualitative discrepancies between observers were resolved by consensus.

Fig. 2.

Quantitative assessment of peritoneal cancer index based on CT (CT-PCI).

(a) Quantitative PCI scheme hand-drawn by the first author (Xia Liu) in GoodNotes v7.0.3(build 3413431.108865178; https://apps.apple.com/app/id1444383602) and adapted from Sugarbaker et al.¹⁷. (b,c) Examples of CT-PCI assessments. CT image (b) shows a peritoneal lesion of 8 mm (white arrow) beside the ascending colon (Region 8), which achieves a score of 2 according to Sugarbaker’s PCI. A fused lesion (c, white arrow) on the pelvic peritoneum (Region 6) achieves a score of 3.

Surgical outcome

Debulking surgery was performed by experienced gynecologists (≥ 20 years) in all patients. The surgical outcome was assessed by the fourth author (15 years’ experience) based on the operation records. R0 resection was defined as no macroscopic residual tumor. Patients in both cohorts were divided into R0 and non-R0 groups according to the surgical outcome.

Statistical analysis

Statistical analyses were conducted using SPSS (version 25.0), MedCalc (version 18.2.1), and R Studio (version 4.1.1). P-value < 0.05 indicated a statistical difference. Continuous variables were expressed as mean ± SD if normally distributed, or as median (lower quartile, upper quartile) if not. Group comparisons for continuous variables used independent-sample t-tests or Mann-Whitney U tests, while categorical variables were analyzed using χ²-tests or Fisher’s exact tests.

Quantitative CT measurements assessed intra-/interobserver agreement via the intraclass correlation coefficient (ICC). ICC > 0.75 indicated good agreement, using first measurements; otherwise, triplicate measurement averages were adopted.

In the training cohort, univariate and multivariate logistic regression analyses identified independent predictors of R0 resection in stage IIB–IV EOC patients. Optimal cut-off values were determined using the Youden Index. A nomogram was developed to predict R0 resection, validated internally via bootstrap resampling (1000 times). The receiver operating characteristic (ROC) and calibration curves were used to assess the nomogram’s performance.

Results

Inter- and intra-observer concordances

The inter- and intra-observer concordances of the overall CT-PCI, upper-abdominal CT-PCI, and the maximum diameter of the primary tumor are shown in Table 1. All ICC values were greater than 0.75, and the first measurements obtained by the first author were taken as the final values for further analyses.

Table 1.

Evaluations of the inter- and intra-observer concordances in the CT-PCI and maximum tumor diameter.

Parameters	Observer 1	Observer 1*	Observer 2	Interobserber ICC (95% CI)	Intraobserver ICC (95% CI)
Overall CT-PCI	13.33 ± 5.07	13.57 ± 5.02	13.60 ± 4.96	0.893 (0.788, 0.948)	0.973 (0.945, 0.987)
Upper-abdominal CT-PCI	3.67 ± 1.95	3.73 ± 2.35	3.630 ± 1.97	0.856 (0.720, 0.929)	0.918 (0.837, 0.960)
Maximum tumor diameter (cm)	8.15 ± 3.70	7.94 ± 4.07	8.13 ± 3.59	0.945 (0.888, 0.973)	0.948 (0.893, 0.975)

CT-PCI, peritoneal cancer index based on computed tomography; ICC, intraclass correlation coefficient; Observer 1, the first measurement by the first observer; Observer 1*, the second measurement by the first observer; Observer 2, the measurement by the second observer.

CT and clinical features of the training and validation cohorts

After debulking surgery, R0 resection was achieved in 66.00% and 61.50% patients in the training and independent validation cohorts, respectively. The CT and clinical features of the training and validation cohorts are shown in Table 2. Except for the histological type and upper-abdominal CT-PCI, no statistical difference was found in the clinical features, CT features, and surgical outcomes between the two cohorts (all P-values > 0.05).

Table 2.

CT and clinical features of the training and independent validation cohorts.

Variables	Training cohort (n = 144)	Validation cohort (n = 65)	P value
Age	54.00 (50.00, 60.00)	54.00 (47.00, 60.00)	0.671
Peripheral hematological parameters
HE4 (pmlol/L)	140.90 (64.50, 398.18)	141.73 (79.50, 264.60 )	0.891
CA-125 (U/mL)	111.05 (31.25, 526.33)	162.20 (35.52, 499.30)	0.765
ALB (g/L)	43.25 (40.35, 46.00)	42.60 (39.90, 45.60)	0.361
NLR	2.68 (2.19, 3.94)	2.94 (1.91, 4.29)	0.757
LMR	3.53 (2.42, 4.73)	3.98 (2.75, 4.87)	0.181
PLR	143.16 (111.97, 230.57)	142.11 (108.33, 236.38)	0.689
SIRI	1.10 (0.71, 2.14)	0.99 (0.54, 1,62)	0.113
Treatment			0.115
PDS	79 (54.90%)	28 (43.10%)
IDS	65 (45.10%)	37 (56.90%)
Histological type			0.001*
Serous carcinoma	126 (87.50%)	60 (92.30%)
Mucinous carcinoma	4 (2.80%)	3 (4.60%)
Clear cell carcinoma	9 (6.30%)	0 (0.00%)
Endometrioid carcinoma	5 (3.50%)	2 (3.10%)
Initial FIGO stage			0.922
IIB	15 (10.4%)	5 (7.70%)
IIIA	10 (6.90%)	5 (7.70%)
IIIB	20 (13.90%)	7 (10.80%)
IIIC	75 (52.10%)	36 (55.40%)
IV	24 (16.70%)	12 (18.50%)
CT features of the primary tumor
Tumor boundary			0.051
Clear	29 (20.10%)	6(9.20%)
Unclear	115 (79.90%)	59(90.80%)
Tumor location			0.150
Unilateral ovary	57 (39.60%)	19 (29.20%)
Bilateral ovaries	87 (60.40%)	46 (70.80%)
Maximum tumor diameter (cm)	6.15 (3.70, 9.86)	4.8 (3.40, 8.50)	0.274
Metastatic lymph node			0.186
Positive	36 (25.00%)	22 (33.80%)
Negative	108 (75.00%)	43 (66.20%)
Overall CT-PCI	8.00 (4.00, 13.00)	11 (6.00, 15.50)	0.060
Upper-abdominal CT-PCI	2.00 (0.00, 4.00)	4.00 (2.00, 8.00)	< 0.001*
Surgical outcome			0.535
R0 resection	95 (66.00%)	40 (61.50%)
Non-R0 resection	49 (34.00%)	25 (38.50%)

HE4, human epididymis protein 4; CA-125, carbohydrate antigen 125; ALB, albumin; NLR, neutrophil-to-lymphocyte ratio; LMR, lymphocyte-to-monocyte ratio；PLR, platelet-to-lymphocyte ratio; SIRI, systemic inflammation response index; PDS, primary debulking surgery; IDS, interval debulking surgery; CT-PCI, peritoneal cancer index based on computed tomography.

Univariate and multivariate logistic regression analyses in the training cohort

In the training cohort, the comparisons of the clinical and CT features between the R0 subgroup and the non-R0 subgroup are shown in Table 3. In order to reduce the model complexity and avoid overfitting, parameters with P < 0.05 were included in the univariate logistic regression analyses, and the results showed that the levels of serum HE4, CA-125, ALB, NLR, PLR, SIRI, treatment, histological type, tumor boundary, tumor location, metastatic lymph node, upper-abdominal CT-PCI, and overall CT-PCI were the predictors of the R0 resection in patients with stage IIB–IV EOC. According to multivariate logistic regression analyses of these parameters, HE4 level (odds ratio (OR) = 1.003, 95%CI: 1.001–1.004), NLR (OR = 1.272, 95%CI: 1.023–1.583), and overall CT-PCI (OR = 1.245, 95% CI: 1.120, 1.385) were the independent predictors of the R0 resection (all P-values < 0.05) (Table 4). Patients with the serum HE4 level ≤ 195.79 pmpl/L, NLR ≤ 3.18, or overall CT-PCI ≤ 9 had a higher probability of R0 resection than those with the serum HE4 level > 195.79 pmol/L, NLR > 3.18, or overall CT-PCI > 9.

Table 3.

Comparisons between the R0 subgroup and non-R0 subgroup in the training cohort.

Variables	R0 subgroup (n = 95)	non-R0 subgroup (n = 49)	P value
Age	54 (50, 59)	54 (50, 66)	0.229
Peripheral hematological parameters
HE4 (pmlol/L)	84.76 (60.10, 183.20)	444.10 (197.90, 901.70)	< 0.001*
CA-125 (U/mL)	68.50 (26.00, 268.00)	367.50 (112.65, 995.50)	< 0.001*
ALB (g/L)	44.50 (41.80, 46.50)	40.60 (36.85, 44.65)	< 0.001*
NLR	2.40 (1.89, 3.06)	3.91 (2.97, 5.57)	< 0.001*
LMR	3.72 (2.91, 5.28)	3.00 (2.15, 4.29)	0.061
PLR	137.27 (106.79, 190.11)	206.92 (129.33, 281.52)	0.001*
SIRI	0.97 (0.61, 1.52)	1.45 (1.01, 3.08)	0.001*
Treatment			0.004*
PDS	44 (46.30%)	35 (71.40%)
IDS	51 (53.70%)	14 (28.60%)
Histological type			0.002
Serous carcinoma	80 (84.20%)	46 (93.90%)
Mucinous carcinoma	1 (1.10%)	3 (6.10%)
Clear cell carcinoma	9 (9.50%)	0 (0.0%)
Endometrioid carcinoma	5 (5.30%)	0 (0.0%)
Initial FIGO stage			0.051
IIB	14 (14.70%)	1 (2.00%)
IIIA	8 (8.40%)	2 (4.10%)
IIIB	14 (14.70%)	6 (12.20%)
IIIC	44 (46.30%)	31 (63.30%)
IV	15 (15.80%)	9 (18.40%)
CT features of the primary tumor
Tumor boundary			0.009*
Clear	23 (24.20%)	6 (12.20%)
Unclear	72 (75.80%)	43 (87.80%)
Tumor location			0.003*
Unilateral ovary	46 (48.40%)	11 (22.40%)
Bilateral ovaries	49 (51.60%)	38 (77.60%)
Maximum tumor diameter (cm)	5.80 (3.40, 9.60)	7 (4.95, 10.55)	0.064
Metastatic lymph node			0.006*
Positive	78 (82.10%)	30 (61.20%)
Negative	17 (17.90%)	19 (38.80%)
Overall CT-PCI	6.00 (3.00, 9.00)	14.00 (10.50,18.50)	< 0.001*
Upper-abdominal CT-PCI	1.00 (0.00, 2.00)	4.00 (3.00, 6.00)	< 0.001*

HE4, human epididymis protein 4; CA-125, carbohydrate antigen 125; ALB, albumin; NLR, neutrophil-to-lymphocyte ratio; LMR, lymphocyte-to-monocyte ratio; PLR, platelet-to-lymphocyte ratio; SIRI, systemic inflammation response index; PDS, primary debulking surgery; IDS, interval debulking surgery; CT-PCI, peritoneal cancer index based on computed tomography.

Table 4.

Univariate and multivariate logistic regression analysis for predicting R0 resection in the training cohort.

Variables	Univariate analysis		Multivariate analysis
	OR (95% CI)	P	OR (95% CI)	P
HE4 (pmlol/L)	1.004 (1.003, 1.006)	< 0.001*	1.003 (1.001, 1.004)	0.012*
CA-125 (U/mL)	1.000 (1.000, 1.001)	0.023*	-	-
ALB (g/L)	0.876 (0.815, 0.942)	< 0.001*	-	-
NLR	1.730 (1.330, 2.251)	< 0.001*	1.272 (1.023,1.583)	0.031*
PLR	1.008 (1.004, 1.012)	< 0.001*	-	-
SIRI	0.966 (0.910, 1.027)	0.268
Treatment	2.898 (1.383, 6.070)	0.005*	-	-
Tumor boundary	2.298 (0.864, 6.076)	0.096
Tumor location	3.243 (1.483, 7.092)	0.003*	-	-
Metastatic lymph node	2.906 (1.335, 6.327)	0.007*	-	-
Overall CT-PCI	1.353 (1.228, 1.491)	< 0.001*	1.245 (1.120, 1.385)	< 0.001*
Upper-abdominal CT-PCI	1.802 (1.466, 2.214)	< 0.001*	-	-

HE4, human epididymis protein 4; CA-125, carbohydrate antigen 125; ALB, albumin; NLR, neutrophil-to-lymphocyte ratio; PLR, platelet-to-lymphocyte ratio; SIRI, systemic inflammation response index; CT-PCI, peritoneal cancer index based on computed tomography; OR, odds ratio.

Development and validation of the nomogram

Based on the independent predictors, a nomogram was developed (Fig. 3). According to the nomogram, patients with a lower total point had a higher probability of R0 resection. The areas under ROC curves (AUCs), sensitivity and specificity for the nomogram to predict R0 resection were 0.908 (95% CI: 0.848–0.95), 89.80% and 77.89%, respectively, in the training cohort and 0.779 (95% CI: 0.659–0.873), 64.00% and 82.50%, respectively, in the independent validation cohort (Fig. 4). The calibration curves showed good agreement between the nomogram predictions and the actual observations in both cohorts (Fig. 5). Furthermore, subgroup analysis was conducted in the validation cohort to evaluate the performance of the nomogram in predicting R0 resection among patients undergoing IDS and PDS. The AUCs were 0.762 (95%CI: 0.594, 0.886) in patients undergoing IDS and 0.806 (95%CI: 0.613, 0.930) in those undergoing PDS (Figure S1).

Fig. 3.

Nomogram for predicting R0 resection in epithelial ovarian cancer patients with stage IIB to IV. Overall CT-PCI, overall peritoneal cancer index based on CT; HE4, human epididymis protein 4; NLR, neutrophil-to-lymphocyte ratio.

Fig. 4.

Receiver operating characteristic curves of the nomogram in the training cohort (a) and independent validation cohort (b).

Fig. 5.

Calibration curves of the nomogram in the training cohort (a) and the independent validation cohort (b).

Discussion

In this study, the overall CT-PCI, upper-abdominal CT-PCI, CT features of the primary tumor and metastatic lymph node, and clinical features before the debulking surgery of patients with stage IIB–IV EOC were analyzed to select independent predictors of R0 resection, and a novel nomogram based on the independent predictors was developed and validated.

This study showed that the overall CT-PCI was one of the independent predictors of R0 resection. Patients with higher overall CT-PCI had a lower probability of R0 resection. As a scoring system, PCI quantitatively assesses peritoneal tumor burden and distribution during surgery¹⁷. Previous studies have found that CT-PCI can evaluate the extent of metastatic tumors in the peritoneal cavity in OC patients noninvasively, and the score size is closely related to the outcome of debulking surgery^18,19. However, the cut-off value of CT-PCI was variable in the reported studies, and the predictive performance should be improved. Llueca et al.¹¹ concluded that PCI > 20 was a major risk factor for R2 resection in patients with advanced OC. Asp et al.¹⁶ demonstrated that the cut-off value of CT-PCI to predict non-R0 resection was 21, with an AUC, sensitivity, and specificity of 0.715, 58.5% and 70.3%, respectively. In OC patients receiving IDS, the cut-off value of CT-PCI to predict R0 resection was 17, with an AUC, sensitivity, and specificity of 0.77, 77%, and 83%, respectively¹⁸. In our study, overall CT-PCI < 9 was associated with a higher probability of R0 resection. The cut-off value of overall CT-PCI was significantly lower than those mentioned above. The potential reasons may be attributed to the methodological variability in the current research, particularly differences in study samples. Our training cohort contained 15 patients with FIGO stage IIB, who generally present with a relatively lower peritoneal tumor burden compared with those with FIGO stage III-IV. Additionally, 65 patients (45.1% of the cohort) had received NACT, which is known to induce tumor regression and decrease tumor burden. As a result, the overall CT-PCI evaluated within two weeks before surgery in patients who underwent NACT may have appeared lower.

In addition, for gynecological oncologists, the technical challenges of cytoreduction in EOC are generally greater in the upper abdomen than in the pelvis. Therefore, the impact of upper-abdominal CT-PCI on achieving R0 resection was evaluated in this study. However, the results of multivariate logistic regression analysis demonstrated that the upper-abdominal CT-PCI was not an independent predictor of R0 resection. The possible reason may be described as follows: Firstly, compared with the overall CT-PCI based on the total peritoneal cavity, the upper-abdominal CT-PCI could not reflect the tumor burden comprehensively. Secondly, a strong correlation likely exists between the upper-abdominal CT-PCI and the overall CT-PCI, which may have led to the exclusion of the upper-abdominal CT-PCI from the multivariate analysis due to its relatively lower predictive value compared with the overall CT-PCI. Serum HE4 and CA 125 are common clinical indicators to assess ovarian tumor burden. Previous studies have shown that the serum HE4 level is better than CA 125 to predict debulking surgery outcomes in EOC patients^20,21. Our study also approved the value of the HE4 level in predicting R0 resection. However, the consensus on the optimal cut-off value of HE4 in predicting R0 resection has not been achieved. Angioli et al.²² showed the cut-off value of HE4 for predicting R0 resection in advanced OC patients was 262pmol/L, whereas, in our study, the cut-off value of HE4 was 195.79 pmol/L. The potential reason may be attributed to the difference in the study samples. In this study, patients receiving NACT followed by IDS were enrolled, and the HE4 levels after NACT before surgery were analyzed, which may be lower than the baseline HE4 levels.

Previous studies have demonstrated that peripheral blood inflammation indexes show a certain clinical application prospect to predict the treatment response, recurrence, and long-term survival of patients with various malignancies^23–25. Multivariate logistic regression showed NLR was the independent predictor. EOC patients with lower NLR have a greater probability of achieving R0 resection.

In this study, the patients underwent PDS or IDS according to the preoperative evaluation by the multiple disciplinary teams, which mainly consisted of gynecologists, radiologists, oncologists, and gastrointestinal and hepatobiliary surgeons. Reported studies illustrated that, in patients who were not suitable for PDS, neoadjuvant chemotherapy before IDS could reduce tumor burden and improve surgical resection²⁶. In this study, the proportion of IDS in the R0 group was higher than that in the non-R0 group. However, according to multivariate logistic regression analyses, the two different treatments were not independent predictors of R0 resection.

Additionally, our study showed that patient age was not associated with the prediction of R0 resection, which was not consistent with the Suidan score system²⁷. In the Suidan score system, age > 60 years is a predictor of the surgical outcome. The Suidan score’s age > 60 predictor may reflect our cohort’s younger median age (54 years in both cohorts). Larger studies across age groups are needed to validate the age impact on R0 resection.

Although the overall CT-PCI, the level of serum HE4, and NLR have been demonstrated to have potential value for predicting R0 resection in EOC patients, previous studies mainly evaluated the predictive value of a single predictor mentioned above^18,28,29. To our knowledge, few studies have combined these factors for prediction. Here, we developed a novel nomogram based on them to predict R0 resection, achieving good performance. As a statistical tool, the nomogram visually displays the probability of R0 resection, making it convenient and easy to understand in clinical practice. Furthermore, subgroup analysis confirmed that the nomogram retained consistent and acceptable predictive performance across both IDS and PDS subgroups in the independent validation cohort, supporting its applicability in predicting R0 resection regardless of treatment strategy.

Our study had several limitations. Firstly, this was a retrospective study with a relatively small sample. A prospective study with a larger sample will be performed to confirm our findings in the future. Secondly, the study population includes not only patients undergoing PDS but also those undergoing IDS; there may be some potential heterogeneity between these two groups. But in clinical practice, the preoperative evaluation of PDS and IDS follows a unified standard, as both aim to determine the likelihood of achieving R0 resection. Although IDS additionally considers chemotherapy response, this is an extension rather than a deviation, and the core assessment principle remains consistent across both approaches. And in this study, the results from univariate and multivariate logistic regression analyses demonstrated that the treatment (PDS or IDS) was not an independent predictor of achieving R0 resection in patients with stage IIB–IV EOC. In addition, because of the limited sample size, the predictive value of post-NACT changes in CT and clinical features for R0 resection was not evaluated in this study. Finally, the CT images were obtained from several different multidetector CT scanners, but all the CT scanners can provide high-quality images to detect the lesions, and the probable influence of different CT scanners on the results of CT features we analyzed in this study could be negligible.

Conclusions

Our study showed that the overall CT-PCI, the level of serum HE4, and NLR before debulking surgery were the independent predictors of R0 resection in stage IIB–IV EOC patients, and a nomogram based on them had reliable predictive performance. These findings are helpful to guide individual treatment.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Data curation, Writing—original draft preparation: [Xia Liu]; Data curation, Formal analysis, Methodology: [Xue-mei Ding], [Qiao-mei Xu], [Ai-ping Wen], [Wei-xiao Luo], [Hui-xin He]; Conceptualization, Methodology, Supervision, Writing—review and editing: [Hai-ying Zhou].

Funding

This work was supported by the Natural Science Foundation of Sichuan Province (Grant Number 2024NSFSC1791).

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

This study was approved by the Medical Ethics Committees of the Affiliated Hospital of North Sichuan Medical College (Institution1, No. 2024ER720-1), the Second School of Clinical Medicine, North Sichuan Medical College (Institution 2, No. 2024 − 176), and the Zigong First People’s Hospital (Institution 3, No. 2024-049). The study was conducted in accordance with the Declaration of Helsinki (2024 version). Due to the retrospective nature of the study, the institutional review boards of the Affiliated Hospital of North Sichuan Medical College, the Second School of Clinical Medicine, North Sichuan Medical College, and the Zigong First People’s Hospital waived the need to obtain informed consent.