Advanced stage, elevated mFIB-4 score, and MMR gene loss as independent predictors of oncological outcomes in endometrioid endometrial cancer: A retrospective observational study

Abstract

This study aimed to evaluate the association between noninvasive liver fibrosis and steatosis scores (including the aspartate transaminase to alanine transaminase ratio to platelet ratio index, aspartate transaminase-platelet ratio index [APRI], Fibrosis-4 index, modified Fibrosis-4 index [mFIB-4], and hepatic steatosis index) as well as ultrasonographic steatosis scores (USS), with histopathological features of endometrioid-type endometrial cancer (EC) and their potential impact on survival outcomes. This retrospective observational study included 415 patients diagnosed with endometrioid-type EC who underwent total abdominal hysterectomy with bilateral salpingo-oophorectomy. Preoperative metabolic scores, including aspartate transaminase to alanine transaminase ratio to platelet ratio index, APRI, Fibrosis-4 index, mFIB-4, and hepatic steatosis index, were calculated from blood samples. USS and the diagnosis of metabolic dysfunction-associated fatty liver disease were based on liver imaging and metabolic criteria. Disease-free survival and overall survival were analyzed; significance was set at P < .05. Analysis showed that patients with body mass index <30 kg/m² had higher rates of deep myometrial invasion (44.8% vs 27.3%, P = .014), serosal invasion (6.9% vs 0.6%, P = .017), lympho-vascular space invasion (34.5% vs 15.8%, P = .002), and MSH6 loss (9.8% vs 1.4%, P = .014) compared to obese patients. USS correlated with age, myometrial invasion depth, and PMS2 loss. Myometrial invasion >1/2 was more frequent in patients with mFIB-4 > 0.19 (37.5% vs 29%, P = .048). Lower APRI (≤0.19) was related to the higher rate of cervical stromal invasion (15.8% vs 9.1%, P = .027), lympho-vascular space invasion (29.7% vs 17.7%, P = .002), adnexal involvement (11.3% vs 5.6%, P = .026), and lymph node metastasis (24.1% vs 12.8%, P = .009). Multivariable analysis showed advanced stage (HR = 5.172, P < .001) and presence of at least one mismatch repair gene defect (HR = 2.936, P = .011) independently predicted poor disease-free survival. Advanced stage (HR = 7.519, P < .001) and a high mFIB-4 score (HR = 2.281, P = .020) independently predicted worse overall survival. In conclusion, advanced stage remains the most significant independent predictor of poor oncological outcomes in endometrioid-type EC. Furthermore, this study highlights the prognostic relevance of metabolic dysfunction in this patient group. Noninvasive liver fibrosis markers (particularly mFIB-4 and APRI) were significantly associated with adverse histopathological features and survival outcomes, suggesting their potential role in risk stratification and prognosis assessment in endometrioid-type EC.

1. Introduction

Rising obesity rates have reshaped global weight trends, shifting many countries from a landscape once dominated by underweight populations to one where obesity now more common.^[1] Obesity can result in various severe long-term health consequences, including cardiovascular diseases, type 2 diabetes mellitus, reproductive disorders, metabolic syndrome, metabolic dysfunction-associated fatty liver disease (MAFLD), cancers, and mental health conditions.^[2–6] Researches are shedding light on the relationship between metabolic dysfunction (including obesity, hyperglycemia, hypertension, and other signs of metabolic abnormalities) and oncogenesis.^[7–9] Recent researches showed that obesity and also metabolic disfunction were related to gynecological cancers including endometrial cancer (EC) and ovarian cancer.^[3,8,9]

EC exhibits one of the strongest links to obesity among all gynecological malignancies.^[3,9,10] The incidence of corpus uterine cancer has risen, making it the second most common gynecologic cancer worldwide.^[11] The staging system for EC underwent a significant revision in 2023, incorporating the latest advances in molecular classification to improve prognostic stratification, and treatment planning, thereby marking a critical step forward in personalized management of the disease.^[12] One of the molecular stratification parameters added in the 2023 revision was the Mismatch Repair Deficient group; however, unlike the well-established prognostic roles of the POLE-mutated and p53 abnormal groups, the impact of the Mismatch Repair Deficient subgroup on prognosis remains less clearly defined in this update.^[12]

Given the rising prevalence of obesity and metabolic syndrome in recent years, this conclusion comes as no surprise. MAFLD, which has seen an increased prevalence alongside obesity, has been reported to be associated with several cancer types including EC.^[13–15] Noninvasive scores or markers are recommended for prediction of MAFLD.^[16] Additionally, these noninvasive scores have been linked to both hepatic and non-hepatic cancers, offering valuable insights into their prognostic factors.^[13–15,17] The Fibrosis-4 index (FIB-4) was initially designed to assess liver fibrosis in individuals with coinfection of human immunodeficiency virus and hepatitis C virus.^[18] The FIB-4 index is a simple, low-cost, and reliable method used to assess liver fibrosis, based on 4 parameters: age, aspartate transaminase (AST), alanine transaminases (ALT), and platelet (PLT) count. In addition to predicting liver fibrosis, the FIB-4 index also plays a role in assessing mortality risk, predicting chronic kidney disease or cardiovascular disease, and evaluating the risk of carcinogenesis.^[18] Furthermore, Crudele et al reported that noninvasive liver fibrosis scores such as aspartate transaminase to alanine transaminase ratio to platelet ratio index (AARPRI), FIB-4, aspartate transaminase-platelet ratio index (APRI), and modified Fibrosis-4 index (mFIB-4) demonstrated a significant predictive value for gynecological cancers, with notable sensitivity, specificity, and odds ratios, suggesting their potential role as risk indicators in women.^[13]

This study aimed to investigate the relationship between noninvasive liver fibrosis and steatosis scores (including AARPRI, APRI, FIB-4, mFIB-4, and hepatic steatosis index [HSI]) as well as ultrasonographic steatosis scores (USS), with histopathological characteristics of endometrioid-type EC, and to explore their potential influence on survival outcomes.

2. Material and methods

We reviewed the records of patients with EC who underwent at least a total abdominal hysterectomy and bilateral salpingo-oophorectomy, and had preoperative blood test results available, at our clinic between September 2019 and December 2024 (n = 478). Data were gathered from the gynecologic oncology clinic’s database and patients’ files. Patients with secondary malignancy, non-endometrioid EC, liver disease, organ transplantation, alcoholism, neoadjuvant chemotherapy, or a preoperative hematological disorder associated with thrombocytosis or thrombocytopenia were excluded (n = 63). This study was approved by the Institutional Ethics Committee of Ankara Bilkent City Hospital (Approval Number: 2024;1/24/768), and all patients provided written informed consent. A total of 415 patients with endometrioid-type EC were included (Fig. 1).

Figure 1.

The flowchart of the study. EC = endometrial cancer.

Body mass index (BMI) was calculated as weight (kg)/height² (m²) and categorized as <30 kg/m² and ≥30 kg/m² (obesity).^[19] Age was classified as <65 or ≥65.^[20] Preoperative USS ranged from 0 (absent steatosis) to 3 (severe steatosis), with USS categorized as <2 or ≥2 according to ultrasonography reports. USS was categorized below (<) score 2, and equal to score 2 and higher (≤). Patients were classified according to the 2023 International Federation of Gynecology and Obstetrics (FIGO) criteria for EC.^[21] Tumor size was based on the largest tumor diameter. Mismatch repair (MMR) gene status that evaluated by immunochemical analysis was recorded according to pathological reports. Median follow-up time was defined as the duration from initial surgery to either the last recorded contact or death. Disease-free survival (DFS) was the interval from surgery to recurrence, and overall survival (OS) was the time from surgery to death due to the disease, excluding perioperative mortality within the first month or last follow-up.

2.1. Blood samples and definitions of scores

Blood samples were collected preoperatively, 1 day before surgery and analyzed for AST, ALT, PLT, and albumin levels (Atellica, Siemens Healthineers, Erlangen, Germany).

The noninvasive scores were calculated using the following formulas:

$$\text{AARPRI}=\frac{(\text{AST/ALT})\times 150}{\text{PLT}}$$

$$\text{APRI}=\frac{(\text{AST/upper limit of normal values (ULN)})\times 100}{\text{PLT}}$$

$$\text{FIB-4}=\frac{\text{Age}\times \text{AST}}{\text{PLT}\times \sqrt{\text{ALT}}}$$

$$\text{mFIB-4}=\frac{10\times \text{Age}\times \text{AST}}{\text{PLT}\times \text{ALT}}$$

$$\text{HSI}=8\times \text{ALT/AST ratio}+\text{BMI}(\text{+2 for diabetes; +2 if female})$$

2.2. Diagnosis of MAFLD

MAFLD is defined as the presence of hepatic steatosis (fatty liver) in abdominal ultrasonography, with one of the following 3 criteria: overweight/obesity (BMI ≥ 23 kg/m²); or type 2 diabetes; or at least 2 metabolic disturbances as follows^[22]: central obesity, prediabetes, hypertension, hyperlipidemia, reduced plasma high density lipoprotein cholesterol, and C-reactive protein levels >2 mg/L. Although the MAFLD criteria include these variables, some of the data for metabolic abnormalities were not available in our dataset. Therefore, data for MAFLD were recorded for patients who had ultrasonographic findings of liver abnormalities and met at least one of the first 2 criteria. If data were missing for the first 2 criteria, it was recorded as missing.

2.3. Statistical analysis

Descriptive statistics were presented as mean ± standard deviation (SD) and median (minimum–maximum) for continuous variables and frequency (percentage) for categorical variables. Differences between groups were assessed using the Pearson Chi-square test or Fisher exact test for categorical variables, as appropriate. Survival analyses were conducted using the Kaplan–Meier method, and differences between survival curves were evaluated with the log-rank test. Cox proportional hazards regression models were utilized to determine independent prognostic factors for DFS and OS. The results of the Cox models were reported as hazard ratios (HRs) with 95% confidence intervals. Statistical significance was set at P < .05. All analyses were performed using SPSS version 30 (IBM Corp., Armonk).

3. Results

This study included 415 patients with endometrioid-type EC with median age of 60 years (range: 28–81 years). The median number of lymph nodes removed was 35 (range: 3–112), and the median tumor size was 40 mm (range: 2–150 mm). Most patients had stage 1a2 disease (48.9%). Median values were 45.77 for HSI, 0.47 for AARPRI, 0.19 for APRI, 0.88 for FIB-4, and 0.19 for mFIB-4 with these indexes categorized based on their median values. The clinical-biochemical and pathological findings for cohort are detailed in Table 1.

Table 1.

Findings of clinic, histopathologic, and complete blood count of all cohorts.

Clinical and histopathologic findings		n (%)
Obesity (kg/m2)
20–24.9		8 (1.9)
Overweight (25–29.9)		50 (12.0)
Obesity class 1 (30–34.9)		66 (15.9)
Obesity class 2 (35–39.9)		37 (8.9)
Obesity class 3 (≥40)		62 (14.9)
Missing		192 (46.3)
Diabetes mellitus
No		73 (17.6)
Yes		140 (33.7)
UK		202 (48.7)
Hypertension
Yes		115 (27.7)
No		98 (23.6)
UK		202 (48.7)
MAFLD
Yes		90 (21.7)
No		118 (28.4)
UK		208 (50.1)
Ultrasonographic steatosis score
<Score 2		227 (54.7)
≥Score 2		78 (18.8)
UK		110 (26.5)
FIGO 2023 stage
Stage 1a1		39 (9.4)
Stage 1a2		203 (48.9)
Stage 1a3		-
Stage 1b		44 (10.6)
Stage 1c		1 (0.2)
Stage 2a		14 (3.4)
Stage 2b		34 (8.2)
Stage 2c		32 (7.7)
Stage 3a1		8 (1.9)
Stage 3a2		3 (0.7)
Stage 3b1		-
Stage 3b2		-
Stage 3c1		13 (3.1)
Stage 3c2		18 (4.3)
Stage 4a		-
Stage 4b		-
Stage 4c		6 (1.4)
FIGO grade
Grade 1		242 (58.3)
Grade 2		124 (29.9)
Grade 3		46 (11.1)
Not reported		3 (0.7)
Depth of myometrial invasion
Only endometrium		42 (10.1)
<1/2		247 (59.5)
≥1/2		126 (30.4)
Uterine serosal invasion
No		405 (97.6)
Yes		10 (2.4)
Cervical stromal invasion
No		370 (89.2)
Yes		45 (10.8)
Lympho-vascular space invasion
Negative		334 (80.5)
Positive		81 (19.5)
Adnexal involvement
Negative		397 (95.7)
Positive		18 (4.3)
Lymphadenectomy
No		155 (37.3)
Yes		260 (62.7)
Lymphatic metastasis*
No		232 (86.9)
Yes		35 (13.1)
Presence of recurrence
No		349 (84.1)
Yes		23 (5.5)
Exitus
No		352 (84.8)
Yes		26 (6.3)
MMR
MLH1 expression loss
Absent		262 (63.1)
Present		70 (16.9)
UK		83 (20.0)
MSH2 expression loss
Absent		334 (80.5)
Present		5 (1.2)
UK		339 (81.7)
MSH6 expression loss
Absent		329 (79.3)
Present		9 (2.7)
UK		77 (18.6)
PMS2 expression loss
Absent		241 (58.1)
Present		98 (23.6)
UK		91 (21.9)
MMR gene expression loss
None of them		241 (58.1)
At least one of them had expression loss		98 (23.6)
UK		76 (18.3)

Findings of indexes	Mean ± SD	Median (min–max)
Albumin	44.02 ± 3.85	44.54 (30.00–53.98)
HSI	47.05 ± 8.93	45.77 (29.29–73.00)
AARPRI	0.56 ± 0.34	0.47 (0.10–3.34)
APRI	0.23 ± 0.15	0.19 (0.05–1.31)
FIB-4	1.03 ± 0.61	0.88 (0.14–6.28)
mFIB-4	0.23 ± 0.16	0.19 (0.03–1.60)

AARPRI = (AST to ALT ratio) to platelet ratio index, APRI = AST-platelet ratio index, FIB-4 = Fibrosis-4 index, FIGO = International Federation of Gynecology and Obstetrics, HSI = hepatic steatosis index, MAFLD = metabolic dysfunction-associated fatty liver disease, mFIB-4 = modified Fibrosis-4 index, MMR = mismatch repair, UK = unknown.

^*Among the patients who underwent lymphadenectomy.

Among patients with endometrioid EC, 25.5% of those with MAFLD had at least 1 MMR gene expression loss, compared to 74.5% without any MMR gene expression loss (P = .01). USS was significantly associated with age, depth of myometrial invasion, and status of PMS2 expression loss. Patients aged ≥65 years were more prevalent in the USS < 2 group (35.7%) than in the USS ≥ 2 group (23.1%) (P = .04). The rate of the deep myometrial invasion (≥2) was 33% and 19.2% among patients with USS < 2 and USS ≥ 2, retrospectively. Deep myometrial invasion (≥2) was observed in 33% of patients with USS < 2 and 19.2% with USS ≥ 2 (P = .021). PMS2 loss was more common in the USS < 2 group (30.6%) than in the USS ≥ 2 group (16.1%) (P = .028). Obesity was linked to deeper myometrial invasion, uterine serosal invasion, status of the lympho-vascular space invasion (LVSI), and MSH6 expression loss. Rates of deep myometrial invasion (44.8% vs 27.3%, P = .014), uterine serosal invasion (6.9% vs 0.6%, P = .017), LVSI (34.5% vs 15.8%, P = .002), and presence of MSH6 expression loss (9.8% vs 1.4%, P = .014) were higher in nonobese patients. Clinical-pathological associations with MAFLD, USS, and BMI are shown in Table 2.

Table 2.

Association between clinicopathological findings and MAFLD, USS, and obesity (univariate analysis).

	MAFLD			USS			Obesity
	Absent	Present	P	<Score 2	≥Score 2	P	<30 kg/m2	≥30 kg/m2	P
Age (yr)
<65	56 (62.2)	83 (70.3)	.218‡	146 (64.3)	60 (76.9)	.040 ‡	34 (58.6)	112 (67.9)	.202‡
≥65	34 (37.8)	35 (29.7)		81 (35.7)	18 (23.1)		24 (41.4)	53 (32.1)
Stage
1–2b	70 (77.8)	99 (83.9)	.263‡	181 (79.7)	66 (84.6)	.343‡	43 (74.1)	139 (84.2)	.087‡
2c–4	20 (22.2)	19 (16.1)		46 (20.3)	12 (15.4)		15 (25.9)	26 (15.8)
Grade
1–2	78 (87.6)	109 (93.2)	.175‡	197 (87.6)	73 (94.8)	.074‡	49 (86.0)	152 (92.7)	.128‡
3	11 (12.4)	8 (6.8)		28 (12.4)	4 (5.2)		8 (14.0)	12 (7.3)
Depth of myometrial invasion
<1/2	57 (63.3)	84 (71.2)	.230‡	152 (67.0)	63 (80.8)	.021 ‡	32 (55.2)	120 (72.7)	.014 ‡
≥1/2	33 (36.7)	34 (28.8)		75 (33.0)	15 (19.2)		26 (44.8)	45 (27.3)
Uterine serosal invasion
Negative	87 (96.7)	118 (100.0)	.079§	223 (98.2)	77 (98.7)	1.000§	54 (93.1)	164 (99.4)	.017 §
Positive	3 (3.3)	0 (0.0)		4 (1.8)	1 (1.3)		4 (6.9)	1 (0.6)
Cervical stromal invasion
Negative	77 (85.6)	107 (90.7)	.252‡	202 (89.0)	71 (91.0)	.612‡	48 (82.8)	149 (90.3)	.124‡
Positive	13 (14.4)	11 (9.3)		25 (11.0)	7 (9.0)		10 (17.2)	16 (9.7)
Lympho-vascular space invasion
Negative	67 (74.4)	97 (82.2)	.175‡	178 (78.4)	67 (85.9)	.151‡	38 (65.5)	139 (84.2)	.002 ‡
Positive	23 (25.6)	21 (17.8)		49 (21.6)	11 (14.1)		20 (34.5)	26 (15.8)
Diameter of tumor (cm)
<4	52 (57.8)	58 (49.2)	.217‡	121 (53.5)	46 (59.7)	.345‡	25 (43.1)	89 (53.9)	.156‡
≥4	38 (42.2)	60 (50.8)		105 (46.5)	31 (40.3)		33 (56.9)	76 (46.1)
Adnexal involvement
Negative	84 (93.3)	115 (97.5)	.179§	215 (94.7)	77 (98.7)	.196§	55 (94.8)	161 (97.6)	.380§
Positive	6 (6.7)	3 (2.5)		12 (5.3)	1 (1.3)		3 (5.2)	4 (2.4)
Lymphatic metastasis*
Negative	61 (89.7)	58 (84.1)	.328‡	127 (87.0)	34 (81.0)	.326‡	37 (88.1)	92 (87.6)	.937‡
Positive	7 (10.3)	11 (15.9)		19 (13.0)	8 (19.0)		5 (11.9)	13 (12.4)
MLH1 expression loss
Absent	49 (66.2)	79 (79.0)	.059‡	131 (73.6)	52 (83.9)	.102‡	39 (79.6)	115 (81.0)	.831‡
Present	25 (33.8)	21 (21.0)		47 (26.4)	10 (16.1)		10 (20.4)	27 (19.0)
MSH2 expression loss
Absent	72 (96.0)	102 (100.0)	.074§	179 (98.4)	63 (100.0)	.571§	49 (96.1)	144 (99.3)	.166§
Present	3 (4.0)	0 (0.0)		3 (1.6)	0 (0.0)		2 (3.9)	1 (0.7)
MSH6 expression loss
Absent	70 (94.6)	101 (99.0)	.163§	175 (96.7)	62 (100.0)	.342§	46 (90.2)	142 (98.6)	.014 §
Present	4 (5.4)	1 (1.0)		6 (3.3)	0 (0.0)		5 (9.8)	2 (1.4)
PMS2 expression loss
Absent	47 (66.2)	76 (78.4)	.079‡	118 (69.4)	52 (83.9)	.028 ‡	34 (73.9)	105 (76.1)	.766‡
Present	24 (33.8)	21 (21.6)		52 (30.6)	10 (16.1)		12 (26.1)	33 (23.9)
MMR gene expression loss status
Absent	42 (56.0)	76 (74.5)	.010 ‡	114 (63.0)	52 (82.5)	.004‡	32 (62.7)	107 (73.8)	.135‡
Present†	33 (44.0)	26 (25.5)		67 (37.0)	11 (17.5)		19 (37.3)	38 (26.2)

Bold values indicate statistical significance at P < .05.

MAFLD = metabolic dysfunction-associated fatty liver disease, MMR = mismatch repair, USS = ultrasonographic steatosis score.

^*Among the patients who underwent lymphadenectomy.

^†At least of the MMR gene defect was present.

^‡Pearson Chi-square test.

^§Fisher exact test.

Age was significantly related to HSI, AARPRI, FIB-4, and mFIB-4. Older patients (≥65 years) were more prevalent in the HSI > 45.77 (42.5% vs 26.2%, P = .012), AARPRI > 0.47 (44.7% vs 24.6%, P < .001), FIB-4 > 0.88 (47.9% vs 21.1%, P < .001), and mFIB-4 > 0.19 (52.2% vs 17.6%, P < .001) groups. mFIB-4 was significantly linked to depth of myometrial invasion, with a higher rate of invasion ≥1/2 in mFIB-4 > 0.19 patients (37.5% vs 29%, P = .048). APRI was significantly associated with cervical stromal invasion (CSI), LVSI, adnexal involvement, and lymphatic metastasis. Lower APRI (≤0.19) was linked to higher rates of CSI (15.8% vs 9.1%, P = .027), LVSI (29.7% vs 17.7%, P = .002), adnexal involvement (11.3% vs 5.6%, P = .026), and lymph node metastasis (24.1% vs 12.8%, P = .009) (Table 3).

Table 3.

The relation between HSI, AARPRI, APRI, FIB-4, mFIB-4, and histopathological factors.

	HSI			AARPRI			APRI			FIB-4			mFIB-4
	≤45.77	>45.77	P	≤0.47	>0.47	P	≤0.19	>0.19	P	≤0.88	>0.88	P	≤0.19	>0.19	P
Age (yr)
<65	79 (73.8)	61 (57.5)	.012 ‡	156 (75.4)	115 (55.3)	<.001 ‡	140 (68.3)	131 (62.4)	.206‡	161 (78.9)	110 (52.1)	<.001 ‡	173 (82.4)	98 (47.8)	<.001 ‡
≥65	28 (26.2)	45 (42.5)		51 (24.6)	93 (44.7)		65 (31.7)	79 (37.6)		43 (21.1)	101 (47.9)		37 (17.6)	107 (52.2)
FIGO 2023 stage
1–2b	93 (71.5)	92 (76.7)	.356‡	179 (75.8)	168 (69.4)	.115‡	173 (70.0)	174 (75.3)	.196‡	175 (74.8)	172 (70.5)	.293‡	184 (77.3)	163 (67.9)	.021‡
2c–4	37 (28.5)	28 (23.3)		57 (24.2)	74 (30.6)		74 (30.0)	57 (24.7)		59 (25.2)	72 (29.5)		54 (22.7)	77 (32.1)
Grade
1–2	98 (92.5)	99 (90.8)	.667‡	186 (89.9)	185 (88.1)	.566‡	183 (88.4)	188 (89.5)	.716‡	184 (90.2)	187 (87.8)	.434‡	192 (91.0)	179 (86.9)	.181‡
3	8 (7.5)	10 (9.2)		21 (10.1)	25 (11.9)		24 (11.6)	22 (10.5)		20 (9.8)	26 (12.2)		19 (9.0)	27 (13.1)
Depth of myometrial invasion
<1/2	89 (68.5)	79 (65.8)	.658‡	160 (67.8)	159 (65.7)	.627‡	163 (66.0)	156 (67.5)	.721‡	161 (68.8)	158 (64.8)	.348‡	169 (71.0)	150 (62.5)	.048 ‡
≥1/2	41 (31.5)	41 (34.2)		76 (32.2)	83 (34.3)		84 (34.0)	75 (32.5)		73 (31.2)	86 (35.2)		69 (29.0)	90 (37.5)
Uterine serosal invasion
Negative	123 (94.6)	116 (96.7)	.429‡	227 (96.2)	230 (95.0)	.541‡	235 (95.1)	222 (96.1)	.608‡	225 (96.2)	232 (95.1)	.568‡	231 (97.1)	226 (94.2)	.123‡
Positive	7 (5.4)	4 (3.3)		9 (3.8)	12 (5.0)		12 (4.9)	9 (3.9)		9 (3.8)	12 (4.9)		7 (2.9)	14 (5.8)
Cervical stromal invasion
Negative	111 (85.4)	103 (85.8)	.920‡	202 (85.6)	216 (89.3)	.227‡	208 (84.2)	210 (90.9)	.027 ‡	198 (84.6)	220 (90.2)	.067‡	204 (85.7)	214 (89.2)	.255‡
Positive	19 (14.6)	17 (14.2)		34 (14.4)	26 (10.7)		39 (15.8)	21 (9.1)		36 (15.4)	24 (9.8)		34 (14.3)	26 (10.8)
Lympho-vascular space invasion
Negative	99 (76.2)	90 (75.6)	.923‡	173 (73.3)	190 (78.8)	.157‡	173 (70.3)	190 (82.3)	.002 ‡	171 (73.1)	192 (79.0)	.129‡	182 (76.5)	181 (75.7)	.850‡
Positive	31 (23.8)	29 (24.4)		63 (26.7)	51 (21.2)		73 (29.7)	41 (17.7)		63 (26.9)	51 (21.0)		56 (23.5)	58 (24.3)
Diameter of tumor (cm)
<4	75 (58.1)	57 (47.5)	.093‡	124 (53.9)	141 (58.5)	.315‡	131 (54.1)	134 (58.5)	.338‡	123 (53.7)	142 (58.7)	.278‡	129 (55.6)	136 (56.9)	.776‡
≥4	54 (41.9)	63 (52.5)		106 (46.1)	100 (41.5)		111 (45.9)	95 (41.5)		106 (46.3)	100 (41.3)		103 (44.4)	103 (43.1)
Adnexal involvement
Negative	119 (91.5)	113 (94.2)	.422‡	216 (91.5)	221 (91.3)	.937‡	219 (88.7)	218 (94.4)	.026 ‡	211 (90.2)	226 (92.6)	.339‡	218 (91.6)	219 (91.3)	.892‡
Positive	11 (8.5)	7 (5.8)		20 (8.5)	21 (8.7)		28 (11.3)	13 (5.6)		23 (9.8)	18 (7.4)		20 (8.4)	21 (8.8)
Lymphatic metastasis*
Negative	82 (86.3)	63 (82.9)	.536‡	117 (77.0)	145 (84.8)	.073‡	132 (75.9)	130 (87.2)	.009 ‡	115 (77.2)	147 (84.5)	.095‡	118 (79.2)	144 (82.8)	.415‡
Positive	13 (13.7)	13 (17.1)		35 (23.0)	26 (15.2)		42 (24.1)	19 (12.8)		34 (22.8)	27 (15.5)		31 (20.8)	30 (17.2)
MLH1 expression loss
Absent	83 (80.6)	82 (80.4)	.973‡	143 (79.4)	147 (79.0)	.923‡	136 (76.0)	154 (82.4)	.133‡	136 (79.1)	154 (79.4)	.942‡	145 (79.2)	145 (79.2)	1.000‡
Present	20 (19.4)	20 (19.6)		37 (20.6)	39 (21.0)		43 (24.0)	33 (17.6)		36 (20.9)	40 (20.6)		38 (20.8)	38 (20.8)
MSH2 expression loss
Absent	101 (97.1)	106 (100.0)	.120§	177 (97.3)	190 (99.5)	.114§	177 (97.3)	190 (99.5)	.114§	168 (97.1)	199 (99.5)	.100§	179 (97.3)	188 (99.5)	.118§
Present	3 (2.9)	0 (0.0)		5 (2.7)	1 (0.5)		5 (2.7)	1 (0.5)		5 (2.9)	1 (0.5)		5 (2.7)	1 (0.5)
MSH6 expression loss
Absent	97 (94.2)	103 (97.2)	.327§	175 (95.6)	185 (97.9)	.218‡	177 (96.7)	183 (96.8)	.955‡	166 (95.4)	194 (98.0)	.160‡	177 (95.7)	183 (97.9)	.233‡
Present	6 (5.8)	3 (2.8)		8 (4.4)	4 (2.1)		6 (3.3)	6 (3.2)		8 (4.6)	4 (2.0)		8 (4.3)	4 (2.1)
PMS2 expression loss
Absent	73 (75.3)	79 (78.2)	.622‡	139 (79.0)	138 (75.8)	.476‡	128 (73.6)	149 (81.0)	.094‡	130 (77.8)	147 (77.0)	.842‡	142 (78.9)	135 (75.8)	.491‡
Present	24 (24.7)	22 (21.8)		37 (21.0)	44 (24.2)		46 (26.4)	35 (19.0)		37 (22.2)	44 (23.0)		38 (21.1)	43 (24.2)
MMR gene expression loss status
Absent	72 (68.6)	79 (74.5)	.338‡	132 (72.1)	136 (71.2)	.842‡	123 (67.2)	145 (75.9)	.062‡	123 (70.7)	145 (72.5)	.698‡	135 (72.6)	133 (70.7)	.694‡
Present†	33 (31.4)	27 (25.5)		51 (27.9)	55 (28.8)		60 (32.8)	46 (24.1)		51 (29.3)	55 (27.5)		51 (27.4)	55 (29.3)

Bold values indicate statistical significance at P < .05.

AARPRI = (AST to ALT ratio) to platelet ratio index, APRI = AST-platelet ratio index, FIB-4 = Fibrosis-4 index, FIGO = International Federation of Gynecology and Obstetrics, HSI = hepatic steatosis index, mFIB-4 = modified Fibrosis-4 index, MMR = mismatch repair.

^*Among the patients who underwent lymphadenectomy.

^†At least of the MMR gene defect was present.

^‡Pearson Chi-square test.

^§Fisher exact test.

Five-year DFS and OS for cohort were 85.1% and 90.3%, respectively. Poor DFS was significantly linked to older age (P = .031), advanced stage (P < .001; Fig. 2), high-grade tumors (P = .034), deep myometrial invasion (≥50%) (P < .001), uterine serosal invasion (P < .001), LVSI (P < .001), CSI (P = .014), a tumor diameter >4 cm (P = .041), adnexal involvement (P < .001), lymphatic metastasis (P < .001), and somatic MMR protein loss [MLH1 loss (P = .037), MSH2 loss (P = .014), PMS2 loss (P = .041), or at least 1 MMR expression loss (P = .004; Fig. 3)] (Table 4). Lower OS was strongly associated with older age (P < .001), advanced stage (P < .001; Fig. 4), deep myometrial invasion (≥50%) (P < .001), uterine serosal invasion (P < .001), LVSI (P < .001), CSI (P < .001), lymphatic metastasis (P < .001), and mFIB-4 > 0.19 (P < .001; Fig. 5) (Table 4).

Figure 2.

Association between stage and disease-free survival (DFS).

Figure 3.

Association between mismatch repair (MMR) gene status and disease free survival (DFS).

Table 4.

Relation between clinical-pathological findings and disease-free survival, and overall survival.

	Disease free survival				Overall survival
	5-year (%)	P value	HR (95% CI)	P value	5-year (%)	P value	HR (95% CI)	P value
Age (yr)
<65	92.2	.031	Reference	.035	92.6	<.001	Reference	<.001
≥65	87.0		2.132 (1.053–4.317)		69.9		3.764 (2.007–7.061)
Obesity
<30 kg/m2	85.5	.234	Reference	.241	83.2	.548	Reference	.550
30 kg/m2≤	91.0		0.573 (0.225–1.455)		84.7		0.776 (0.337–1.784)
FIGO 2023 stage
1–2b	95.8	<.001	Reference	<.001	93.0	<.001	Reference	<.001
2c–4	70.4		8.327 (3.908–17.744)		60.0		8.633 (4.488–16.608)
Grade
1–2	94.2	.034	Reference	.043	91.7	.076	Reference	.085
3	86.2		2.843 (1.033–7.825)		82.6		2.390 (0.885–6.450)
DMI
<1/2	95.3	<.001	Reference	<.001	93.8	<.001	Reference	<.001
≥1/2	78.0		4.719 (2.260–9.855)		67.0		5.629 (2.935–10.798)
Uterine serosal invasion
Negative	91.5	<.001	Reference	<.001	87.9	<.001	Reference	<.001
Positive	35.1		10.115 (3.789–27.007)		16.3		8.683 (4.121–18.296)
Cervical stromal invasion
Negative	91.4	.014	Reference	.019	87.3	.002	Reference	.003
Positive	80.7		2.743 (1.181–6.375)		69.2		2.824 (1.423–5.607)
LVSI
Negative	95.6	<.001	Reference	<.001	92.7	<.001	Reference	<.001
Positive	68.5		8.277 (3.959–17.305)		58.6		7.656 (4.084–14.352)
Diameter of tumor (cm)
<4	93.4	.041	Reference	.046	88.3	.150	Reference	.154
≥4	87.2		2.106 (1.014–4.373)		80.2		1.553 (0.847–2.847)
Adnexal inv.
Negative	93.8	<.001	Reference	<.001	88.1	<.001	Reference	<.001
Positive	22.1		11.738 (5.599–24.607)		39.2		5.383 (2.683–10.798)
Lym. Met.*
Negative	93.3	<.001	Reference	<.001	89.4	<.001	Reference	<.001
Positive	55.4		6.668 (3.030–14.671)		42.3		8.594 (4.450–16.596)
MLH1 exp. loss
Absent	93.1	.037	Reference	.044	87.1	.511	Reference	.514
Present	83.8		2.370 (1.025–5.478)		86.6		0.724 (0.275–1.908)
MSH2 exp. loss
Absent	91.3	.014	Reference	.027	85.8	.456	Reference	.620
Present	62.5		5.105 (1.199–21.731)		100.0		0.048 (0.000–7623.810)
MSH6 exp. loss
Absent	90.7	.955	Reference	.955	85.3	.265	Reference	.462
Present	88.9		1.059 (0.143–7.847)		100.0		0.047 (0.000–164.791)
PMS2 exp. loss
Absent	93.2	.041	Reference	.048	84.8	.069	Reference	.083
Present	84.0		2.362 (1.009–5.529)		91.0		0.346 (0.105–1.148)
MMR gene exp. loss status
Absent	94.5	.004	Reference	.007	85.6	.176	Reference	.184
Present†	83.6		3.135 (1.374–7.153)		89.0		0.544 (0.222–1.335)
HSI
≤45.77	90.0	.626	Reference	.628	84.4	.713	Reference	.713
>45.77	91.5		0.788 (0.300–2.070)		87.9		0.850 (0.358–2.019)
AARPRI
≤0.47	89.1	.834	Reference	.834	89.8	.068	Reference	.072
>0.47	91.9		0.927 (0.458–1.877)		79.3		1.767 (0.951–3.286)
APRI
≤0.19	89.5	.792	Reference	.793	82.3	.156	Reference	.160
>0.19	91.0		0.909 (0.448–1.846)		88.9		0.637 (0.340–1.194)
FIB-4
≤0.88	90.2	.832	Reference	.832	89.6	.117	Reference	.122
>0.88	90.4		1.079 (0.533–2.185)		79.0		1.623 (0.879–2.997)
mFIB-4
≤0.19	91.0	.320	Reference	.323	93.0	<.001	Reference	.001
0.19	90.2		1.434 (0.702–2.930)		75.7		3.169 (1.594–6.302)
MAFLD
Absent	84.7	.449	Reference	.452	83.7	.383	Reference	.386
Present	89.1		0.731 (0.322–1.657)		85.2		0.706 (0.322–1.550)
USS
<Score 2	88.5	.671	Reference	.672	85.0	.811	Reference	.811
≥Score 2	90.1		0.822 (0.332–2.037)		86.2		1.098 (0.510–2.363)

Bold values indicate statistical significance at P < .05.

AARPRI = (AST to ALT ratio) to platelet ratio index, APRI = AST-platelet ratio index, CI = confidence interval, DMI = depth of myometrial invasion, exp. = expression, FIB-4 = Fibrosis-4 index, FIGO = International Federation of Gynecology and Obstetrics, HR = hazard ratio, HSI = hepatic steatosis index, inv. = involvement, LVSI = lympho-vascular involvement, Lym. Met. = lymphatic metastasis, MAFLD = metabolic dysfunction-associated fatty liver disease, mFIB-4 = modified Fibrosis-4 index, MMR = mismatch repair, USS = ultrasonographic steatosis score.

^*Among the patients who underwent lymphadenectomy.

^†At least of the MMR gene defect was present.

Figure 4.

Association between stage and overall survival (OS).

Figure 5.

Association between mFIB-4 and overall survival (OS). mFIB-4 = modified Fibrosis-4 index.

Advanced stage (2c–4) significantly increased the risk of recurrence (HR = 5.172, P < .001) and death (HR = 7.519, P < .001) compared to lower stages (1–2b). Patients with at least one of the MMR gene defects had a higher recurrence risk (HR = 2.936, P = .011). An mFIB-4 score >0.19 doubled the death risk (HR = 2.281, P = .020). Disease stage and MMR gene defects were independent predictors of DFS, while stage and mFIB-4 score predicted OS (Table 5).

Table 5.

Multivariable Cox regression.

	HR (95% CI)	P value
Disease free survival
Stage
1–2b	Reference	<.001
2c–4	5.172 (2.263–11.823)
MMR gene expression loss status
Absent	Reference	.011
Present*	2.936 (1.284–6.717))
Overall survival
Stage
1–2b	Reference	<.001
2c–4	7.519 (3.879–14.574)
mFIB-4
≤0.19	Reference	.020
>0.19	2.281 (1.138–4.572)

Bold values indicate statistical significance at P < .05.

CI = confidence interval, HR = hazard ratio, mFIB-4 = modified Fibrosis-4 index, MMR = mismatch repair.

^*At least of the MMR gene defect was present.

4. Discussion

Our study found that advanced stage strongly predicted poorer DFS and OS in endometrioid EC. Elevated mFIB-4 that linked to fibrosis, steatosis, or inflammation, was independently correlated with worse OS and was significantly associated only with deep myometrial invasion. Lower APRI was significantly related to CSI, LVSI, adnexal involvement, and lymphatic metastasis. Additionally, MMR deficiency emerged as a significant adverse factor for DFS.

Endometrioid type EC is one of the most associated cancers with obesity, hyperlipidemia, hyperestrogenism, and also metabolic syndrome development.^{[3,8–10,23,24]} Karkia et al reported that all components of metabolic syndrome independently increase EC development risk, regardless of menopausal status.^[10] Furthermore, the risk of developing EC demonstrated a significant temporal increase, proportionate to the number of metabolic syndrome components present. Based on their findings, they suggested that metabolic syndrome could serve as a useful predictor for EC.^[10] Obesity is strongly related to the fatty liver which acts as a source of hyperlipidemia that is major components of metabolic syndrome.^[5,7,9,25] The one of strongest independent predictors among metabolic syndrome parameters for EC risk was asserted as hypertriglyceridemia for both premenopausal and postmenopausal women.^[10]

While metabolic syndrome includes together several metabolic disorders, nonalcoholic fatty liver disease (NAFLD) stands apart as a condition where fat quietly builds up in the liver, independent of alcohol use.^[25] MAFLD is a newer term that aims to better reflect the role of metabolic dysfunction in the development of fatty liver disease.^[22] It is a more specific classification that emphasizes the association between fatty liver and metabolic abnormalities, such as obesity, insulin resistance, and type 2 diabetes.^[22] A positive association has been observed between MAFLD and endometrial thickening in postmenopausal women by Wei et al.^[26] When endometrial thickness ≥5 mm in postmenopausal women, the risk of both ultrasonographic hepatosteatosis and MAFLD is higher.^[26] In light of these findings, recent years have witnessed a growing body of research exploring the potential association between NAFLD or MAFLD, and EC, driven by the shared metabolic disturbances underlying both conditions. Previous studies have yielded mixed results between NAFLD and gynecological cancer risk.^[27–34] The studies that concluded no significant association included all gynecological cancers collectively and did not specifically assess uterine cancer on its own.^[29,32,33] In contrast, Mantovani et al’s meta-analysis suggested that NAFLD is related to an increased risk of gynecological cancers.^[30] Furthermore, Allen et al, Björkström et al, and Park et al each highlighted a notable connection between NAFLD and uterine cancer, indicating that women affected by NAFLD carried a heightened risk of developing uterine cancer in young, middle and older age groups.^[27,28,31]

Clinical guidelines recommend noninvasive scores in diagnosis and management of MAFLD, although liver biopsy remains the gold standard for diagnosis liver fibrosis.^[16,35] Furthermore, these noninvasive scores may use as a predictor for morbidity and risk of malignancy development in NAFLD patients.^[13,17,36] Although noninvasive scores were also found in association with extrahepatic cancers development, Peleg et al found no association with gynecological cancer.^[17] Crudele et al evaluated the relation between noninvasive liver fibrosis scores and cancer development in a cohort including metabolic women.^[13] They concluded that high noninvasive scores including AARPRI, APRI, FIB-4, and mFIB-4 scores were significantly associated with an increased gynecological cancers development risk, particularly for both uterine and ovarian cancer separately, with AARPRI showing the strongest link.^[13] However, when viewed from the standpoint of EC, it remains unclear whether if these scores predict histopathological characteristics or survival outcomes. Lin et al reported that MAFLD was significantly associated with CSI in EC, showing a negative correlation.^[14] In both the overall EC and specifically endometrioid EC groups, the prevalence of MAFLD and ultrasound-detected fatty liver was higher in those without CSI, with no significant link in type II EC.^[14] After adjusting for BMI and age, MAFLD or elevated HSI levels was significantly related to presence of CSI in nondiabetic endometrioid EC group.^[14] This finding is noteworthy because, although obesity is a well-established risk factor for the development of endometrioid EC, it is also commonly associated with early-stage disease and less aggressive tumor findings,^[37] an observation reinforced by our study results. By adjusting for both BMI and T2DM, the study was able to isolate the independent effect of metabolic liver dysfunction.^[14] The study results highlighted that MAFLD is an independent risk factor for CSI in nondiabetic patients with endometrioid EC and suggested that HSI may have predictive biomarker in this subgroup.^[14] However, since the HSI includes BMI and diabetes in its formula, excluding these parameters in analysis may create confusion regarding the true representation of HSI. In our study, the finding of no association between HSI and CSI or other pathological parameters, may be attributed to this issue. Instead, other scores were come into prominence. According to our study findings, a potential association between mFIB-4 index and tumor aggressiveness was shown. Specifically, mFIB-4 was significantly associated with the depth of myometrial invasion, as patients with mFIB-4 > 0.19 had a higher frequency of deep myometrial invasion (>50%) compared to those with lower scores. This may indicate that elevated mFIB-4, reflecting underlying systemic metabolic or inflammatory conditions, could be linked to more invasive tumor behavior. Interestingly, lower APRI (<0.19) was unexpectedly linked to more aggressive histopathological features, including increased rates of CSI, LVSI, adnexal involvement, and lymph node metastasis. This counterintuitive relationship suggests that the interplay between liver fibrosis markers and tumor biology may be more complex than previously thought. It is possible that in this context, lower APRI scores could reflect underlying biological processes unrelated to fibrosis severity but linked to tumor invasiveness.

mFIB-4 can discriminate steatohepatitis in early stages, making it a preferred first-line tool for evaluating liver fibrosis and MAFLD.^[34] It has improved diagnostic accuracy when compared across various serum markers.^[38] Elevated mFIB-4 scores predict not only for hepatocellular carcinoma but also increased risk of cardiovascular disease, overall mortality, and extrahepatic malignancies.^[34] In our study, advanced stage and high mFIB 4 was found as independent poor prognostic factor for OS among endometrioid EC patients. Advanced stage of disease was significantly associated with increased risks of mortality, with approximately a 7-fold higher risk of death compared to early-stage disease. Additionally, an mFIB-4 score >0.19 was associated with a doubled risk of death, suggesting that this noninvasive marker of liver fibrosis may have potential prognostic value in EC.

The differing components of each score help explain their associations. APRI includes PLT count, which may be artificially low in liver disease but elevated in many cancers due to tumor-related thrombocytosis. Thrombocytosis plays an active role in carcinogenesis by promoting angiogenesis, tumor invasion, and creating tumor microenvironment.^[39] Therefore, a low APRI in cancer may reflect thrombocytosis rather than low fibrosis, paradoxically correlating with worse histopathological features. In contrast, mFIB-4 incorporates additional elements such as age, AST, and ALT which may reflect systemic inflammation, hepatic stress, or frailty. Importantly, in many malignancies, adverse outcomes are associated with disproportionately elevated AST compared to ALT,^[40,41] contributing to higher mFIB-4 score and its link to poor prognosis.

Several studies have investigated the clinical risk profiles of endometrioid-type EC, comparing those with and without MMR deficiency.^[42–44] Both lower and higher BMI have been linked to MMR-deficient EC in the literature, though in our study, only MSH6 loss found as related to obesity.^[42–44] Among the metabolic diseases, Nagle et al reported strong association between diabetes and somatic MMR-deficient EC patients compared to MMR-intact patients.^[44] To the best of our knowledge, no study has evaluated the relationship between USS, MAFLD status, or noninvasive scores and somatic MMR status in endometrioid-type EC. In our study, PMS2 loss was less frequent in patients with USS ≥ 2, and presence of at least one of MMR deficiency was significantly less common in those with MAFLD. None of the noninvasive indices showed a significant association with MMR status. According to the guidelines, MMR status is positioned prognostically between POLE-mutated and p53-abnormal subtypes in EC, based on the molecular classification.^[21,23] However, the prognostic value of MMR deficiency in EC remains highly variable, with some studies reporting conflicting results.^[23,45,46] In our study, advanced stage and the presence of at least one somatic MMR deficiency were identified as independent adverse risk factors for DFS. Advanced disease stage was also significantly associated with increased risks of recurrence, with approximately a 5-fold higher risk of recurrence compared to early-stages. Patients with at least 1 MMR gene defect showed nearly a 3-fold increased risk of recurrence, underscoring the prognostic importance of molecular alterations in endometrioid type-EC.

Our study includes large sample size for endometrioid EC that is a strong and diverse sample for comprehensive analysis. To the best of our knowledge, this study is the first to comprehensively evaluate a range of noninvasive scores (including AARPRI, FIB-4, mFIB-4, APRI, and HIS) coupled with ultrasound-based steatosis to investigate the intricate relationship between metabolic dysfunction and histopathological features, as well as between metabolic dysfunction and survival in endometrioid-type EC. The major limitation is retrospective design of the study. One of the other limitations is being single-center study. In our study, MAFLD data were recorded for patients who had ultrasonographic evidence of hepatic steatosis and met at least one of the first 2 diagnostic criteria (overweight/obesity or type 2 diabetes mellitus). If data for these criteria were missing, the case was recorded as missing. However, due to the unavailability of detailed data on other MAFLD criteria, we were unable to apply the full diagnostic framework for MAFLD. Consequently, our ability to fully characterize MAFLD and also to comprehensively assess the relationship between MAFLD and both histopathological features and survival outcomes in endometrioid-type EC may be limited, potentially leading to an underestimation of this association.

In conclusion, the advance stage is still the major independent factor for oncological outcomes in endometrioid type EC. Additionally, this study emphasizes the important relationship between metabolic dysfunction and the oncological outcomes of endometrioid type EC, with a particular focus on the predictive role of noninvasive liver fibrosis markers such as mFIB-4 and APRI in forecasting disease outcomes. Importantly, a low APRI score is found to predict adverse histopathological features, adding an additional layer of clinical relevance. Literature has shown that metabolic disorders contribute to the development and increased risk of EC. Our study reveals that elevated mFIB-4 in EC patients is associated with poor OS, suggesting that efforts to address, raise awareness of, and treat metabolic issues during the follow-up period may be crucial for improving patient outcomes. Presence of at least 1 MMR gene defect was associated with a nearly 3-fold increased risk of recurrence, highlighting the critical prognostic role of molecular alterations in endometrioid type EC management. These findings suggest that integrating molecular profiling with noninvasive fibrosis markers may have the potential to enhance risk stratification and support personalized treatment strategies in patients with endometrioid type EC, although further research is needed to confirm this possibility. These results highlight the need for large-scale, prospective, multicenter studies to confirm and expand these findings.

Author contributions

Conceptualization: Gunsu Kimyon Comert, Zeliha Firat Cuylan.

Data curation: Neslihan Bayramoglu, Eda Kayali, Emel Çevikkan.

Formal analysis: İrem Kar.

Investigation: Gunsu Kimyon Comert, Neslihan Bayramoglu, Eda Kayali, Derya Ari.

Methodology: Gunsu Kimyon Comert, Derya Ari.

Supervision: Meral Akdogan Kayhan, Taner Turan.

Writing – original draft: Gunsu Kimyon Comert.

Writing – review & editing: Zeliha Firat Cuylan, Meral Akdogan Kayhan.