Evaluation of histological tumor necrosis in pancreatic ductal adenocarcinoma via the quantitative parameters from enhanced computed tomography and its relationship with tumor prognosis

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease with a 5-year overall survival (OS) rate of 11% (1). Although surgical resection is the only option that allows potential long-term survival (2), up to 80% of patients experience local tumor recurrence or distant metastasis after surgery, even in the early stages (3-6). Therefore, the preoperative detection of prognostic factors is important to developing personalized treatment strategies for patients with PDAC.

Previous studies have reported that in patients with PDAC, several histological findings, such as tumor grade, vascular invasion, neural invasion, and lymphatic invasion, are biomarkers for predicting postoperative prognosis (7-11), but these are difficult to detect noninvasively on preoperative imaging. Histological tumor necrosis (HTN) occurs in more than 60% of PDAC cases and is closely correlated with a poor prognosis (12,13). The noninvasive and preoperative prediction of HTN may facilitate more efficient treatment strategies to be devised preoperatively. Contrast-enhanced computed tomography (CE-CT) is considered to be the standard modality for depicting, staging, and resecting PDAC (14). Typical PDAC shows hypoenhancement on CE-CT due to the presence of abundant intratumoral fibrosis in the stroma (15,16). However, some parts of the tumor show almost no enhancement or obvious hypoattenuation relative to the surrounding enhanced portion of the tumor on CE-CT. These are known as poorly enhanced areas (PEAs) and are particularly evident during the portal venous phase. Theoretically, PEAs observed on CE-CT may be used as indicators of HTN (17). However, PEAs have been inconsistently defined across various studies, have been subjectively evaluated by a radiologist (who may have difficulty in distinguishing PEAS from the weak enhancement of the tumor), or have been determined under only a single CT threshold standard; the clinical value of PEAs, therefore, remains controversial.

Thus, the purpose of this study was to quantitatively calculate the enhancement difference (delta) between PDAC in the portal venous phase and unenhanced phase based on the automatically registered pancreatic CE-CT image, assess the diagnostic accuracy of delta for the detection of HTN in resectable PDAC, and determine the relationship between computed tomography (CT)-defined necrosis and the prognosis of patients with resectable PDAC. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-905/rc).

Methods

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Ethical approval was received from the Medical Ethics Committee of Shengjing Hospital of China Medical University (ethical approval No. 2021ps830k). The requirement for informed consent was waived due to the retrospective nature of the analysis. All collaborating institutions received complete study documentation and furnished formal letters of approval.

Patients

For this retrospective analysis, we selected patients from three independent sources: our hospital (Shengjing Hospital of China Medical University) (n=966; January 2010 to December 2020) and two external hospitals: Guangdong Provincial People’s Hospital and Tianjin Tumor Hospital (n=150; January 2018 to December 2020). A total of 1,649 consecutive patients with pathologically confirmed margin-negative (R0) PDAC underwent some type of curative pancreatectomy. The inclusion criteria were as follows: (I) preoperative CE-CT of the pancreas completed within 1 month before surgery; (II) no history of neoadjuvant treatment before surgery; and (III) completion of more than 6 months of follow-up after surgery. Figure 1 provides a flowchart of participant inclusion in this study.

Figure 1 Flowchart of participant inclusion in the study. Institute A: Shengjing Hospital of China Medical University; Institute B/C: Guangdong Provincial People’s Hospital/Tianjin Tumor Hospital. CE-CT, contrast-enhanced computed tomography; PDAC, pancreatic ductal adenocarcinoma.

Image acquisition

Preoperative CE-CT of the pancreas was performed in routine fashion during the unenhanced, pancreatic parenchymal, and portal venous phases. To image the patients in our hospital, we used two multidetector CT scanners: a 320-channel scanner (Aquilion ONE 640; Canon Medical Systems, Otawara, Japan) and a 256-channel scanner (Brilliance 128; Philips Medical Systems, Amsterdam, the Netherlands). At the external hospitals, one 256-channel scanner (Brilliance 128; Philips Medical Systems) was used to image patients. Nonionic contrast material (iopromide solution with an iodine concentration of 370 mg/mL) was delivered to each patient with a power injector (3.0–5.0 mL/s) via an 18-gauge catheter placed in the antecubital vein. The average time delays to imaging from the injection of contrast to the starting points of the pancreatic parenchymal and portal venous phase were 35–45 and 60–70 seconds, respectively, which is in accordance with the latest National Comprehensive Cancer Network stipulations (18).

CT image analysis

The dense energy displacement sampling (DEEDS) registration method has superior performance in the volume registration of abdominal CT images; therefore, we registered all CE-CT images using the DEEDS registration method (Figure S1). In each center, patients underwent the pancreatic parenchymal phase as the fixed phase, with the images being registered in the unenhanced phase and portal venous phase. This ensured voxel-wise correspondence across phases, eliminating misalignment due to respiratory motion or patient positioning. Tumor segmentation was performed with three-dimensional no new UNet (3D-nnUNet) on the registered isotropic volume (1×1×1 mm³), generating a three-dimensional (3D) mask encompassing the entire tumor volume. The methods for the above-mentioned automatic registration and tumor recognition segmentation model were published in the Annals of Surgery in 2022 (19). Attenuation differences (delta) were calculated for every voxel within the 3D tumor mask between the registered portal venous and unenhanced phases via MATLAB (MathWorks, Natick, MA, USA). The final delta value per tumor represented the mean of all voxel-wise differences. This approach enabled whole-tumor volumetric analysis rather than single-slice measurements. We selected the delta threshold for CT-defined necrosis within the range of 10–30 Hounsfield units (HU) in accordance with previous studies (17,20,21). In addition, by combining the low-density area with clear boundaries on CT and the area of high signal intensity on the magnetic resonance T2-weighted image, we were able to exclude cysts, dilated pancreatic ducts, and other factors. In addition, the subjective evaluation was conducted by two radiologists with 16 and 10 years of experience in pancreatic imaging, who independently and blindly assessed the presence of tumor necrosis on CT images without knowing the pathological results.

Clinical and pathological data collection

Resected PDAC specimens were fixed in 10% formalin and cut into serial 5-mm-thick slices. Subsequently, 3-µm-thick sections were stained with hematoxylin and eosin (H&E). All H&E-stained sections were examined under a batch slide scanner (NanoZoomer 2.0-RS, Hamamatsu Photonics, Hamamatsu City, Japan). The pathological evaluations of all specimens were reviewed by two experienced pathologists (with 9 and 14 years of experience in pancreatic pathology, respectively) who were blinded to clinical and radiologic patient data. Gross necrosis was considered to be the presence of amorphous pale yellow or pale lesions (22). HTN was defined as confluent cell death in invasive areas visible at an objective lens magnification of 4×, with the disappearance of nuclei but cell contours being retained (12,13).

We extracted demographic and clinical data from the digital medical records, while pathology reports provided other relevant data, such as tumor location, tumor size, histological grade, and perineural invasion. Each patient was followed up after surgery by CE-CT or magnetic resonance imaging studies and via the monitoring of laboratory profiles every 3 to 6 months (18). Disease-free survival (DFS) was defined as the interval from the date of surgery to tumor recurrence or the last follow-up. OS was defined as the period from the date of surgery to patient death or the most recent outpatient-based follow-up. The last follow-up date was December 31, 2021.

Statistical analysis

Continuous variables are presented as the mean ± standard deviation or as the median and interquartile range, with categorical variables reported as counts with percentages. We compared categorical variables using the χ² or Fisher exact test. The Student t-test and Mann-Whitney test were used to compare continuous variables following normal and nonnormal distributions, respectively. We used the Shapiro-Wilk test to determine the normality of distribution. Interobserver agreement for the presence of HTN was indicated by Cohen’s kappa coefficient (κ), which was interpreted as follows: 0.10–0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.00, excellent. We evaluated the correlation between CT-defined necrosis and HTN using the Spearman rank correlation coefficient. Receiver operating characteristic (ROC) curve analysis was used to determine the diagnostic performance of delta for HTN. The areas under the ROC curve (AUCs) were compared with the DeLong test (23).

We established two multivariate Cox models, one based on preoperative features and the other on pre- and postoperative features, for predicting DFS and OS, respectively. Univariate and multivariate analyses of factors related to DFS and OS were conducted with a Cox proportional hazards model. Factors with a P value less than 0.05 on univariate analysis were included in the multivariate analysis. The hazard ratio (HR) is presented with the 95% confidence interval (CI). Estimates of DFS and OS were generated via the Kaplan-Meier method with log-rank analysis.

All computations were powered by SPSS 25.0 (IBM Corp, Armonk, NY, USA) and MedCalc v.20.1.0 (MedCalc Software, Ostend, Belgium). Statistical significance was set at a P value <0.05.

Results

Clinicopathological and radiologic characteristics

Figure 1 displays a schematic of the patient selection. From the three participating institutions, patients with no CE-CT within 1 month before surgery (n=151), suboptimal image quality (n=72), residual disease (R1 or R2 resection; n=145), a history of neoadjuvant therapy or extrapancreatic cancer (n=69), or a lack of clinical or follow-up data (n=96) were excluded.

Our hospital included 966 patients (mean age 61.3±10.4 years; 546 males). External hospitals included 150 patients (mean age 64.5±10.1 years; 92 males). Table 1 summarizes the demographic and clinical characteristics.

Diagnostic performance of CT-defined necrosis

Our review of the histopathology of 328 patients with PDAC (Figures 2,3) indicated that 205 patients developed HTN (57.2%). Interobserver agreement for the presence of HTN was good (κ=0.90, 95% CI: 0.85–0.95). We noted a significant correlation between CT-defined necrosis and HTN (r=0.78, P<0.001). For diagnosis of HTN, the cutoff of 15 HU performed best (AUC =0.93, 95% CI: 0.91–0.96, P<0.05) and was superior to other CT value difference ranges (10–30 HU) (AUC =0.75–0.87, P<0.05). CT-defined necrosis also outperformed the two radiologists’ subjective diagnosis of HTN (radiologist 1: AUC =0.78, 95% CI: 0.73–0.82; radiologist 2: AUC =0.74, 95% CI: 0.69–0.79; DeLong test: both P values <0.001; Table S1). CT-defined necrosis was positively associated with differentiation (r=0.18, P=0.001) and tumor size (r=0.32, P<0.001).

Figure 2 CT-defined necrosis and HTN consistency of PDAC (2.2 cm at head) in a 55-year-old woman (the arrows in each picture indicate where the tumor necrosis was located). (A) Image of PDAC on CE-CT in the unenhanced phase. (B) Image of PDAC on CE-CT in the pancreatic parenchymal phase. (C) Image of PDAC on CE-CT in the portal venous phase. The subtraction image in the lower-right corner shows tumor segmentation with a central necrotic area in white. (D,E) The panoramic section scanned at 40× magnification clearly delineates the tumor location, with the 400× magnification image of H&E staining sections indicating where the necrosis occurred. (F) Image of the necrotic area in the gross specimen after pancreatectomy. CE-CT, contrast-enhanced computed tomography; CT, computed tomography; H&E, hematoxylin and eosin; HTN, histological tumor necrosis; PDAC, pancreatic adenocarcinoma.

Figure 3 CT-defined necrosis and HTN consistency of PDAC (1.5 cm at head) in a 61-year-old woman (the arrows in each picture indicate the location of the tumor). (A) Image of PDAC on CE-CT in the unenhanced phase. (B) Image of PDAC on CE-CT in the pancreatic parenchymal phase. (C) Image of PDAC on CE-CT in the portal venous phase. (D,E) Panoramic scanning (magnification, 40×), with the 400× magnification image of H&E staining sections indicating where the tumor occurred. (F) Image of the PDAC area in gross specimen after pancreatectomy. CE-CT, contrast-enhanced computed tomography; CT, computed tomography; H&E, hematoxylin and eosin; HTN, histological tumor necrosis; PDAC, pancreatic adenocarcinoma.

Analysis of the factors for predicting DFS and OS

We conducted univariate and multivariate Cox regression analyses in all patients, as shown in Tables 2-4. Of the 966 patients in our hospital, 724 (74.9%) experienced tumor recurrence during follow-up. The median DFS was 20 months, and the cumulative DFS rates at 1, 2, and 3 years were 68.1%, 39.7%, and 21.3%, respectively. When using multivariate preoperative factors (model 1 in Tables 2,3), we found that CT-defined necrosis was an independent influencing factor of DFS in patients in our hospital (HR =2.33, 95% CI: 1.98–2.76, P<0.001) and the patients in external hospitals (HR =2.82, 95% CI: 1.88–4.26, P<0.001). When we used both pre- and postoperative factors (model 2 in Tables 2,3), CT-defined necrosis was also independently associated with unfavorable DFS in all patients (our hospital: HR =2.22, 95% CI: 1.88–2.62, P<0.001; external hospitals: HR =2.31, 95% CI: 1.51–3.51, P<0.001).

Of the 966 patients in our hospital, 654 (67.7%) died during follow-up. The median OS was 27 months, and the cumulative OS rates at 1, 2, and 3 years were 88.4%, 57.8%, and 34.6%, respectively. In the preoperative setting (model 1 in Tables 4,5) for OS, CT-defined necrosis was an independent prognostic factor in our hospital (HR =1.79, 95% CI: 1.50–2.13, P<0.001). Similar results were also found in external hospitals for CT-defined necrosis (HR =1.92, 95% CI: 1.30–2.84, P<0.001). When both pre- and postoperative factors were considered (model 2 in Table 4), we found that CT-defined necrosis was independently associated with poor OS in our hospital (HR =1.70, 95% CI: 1.43–2.02, P=0.001); however, this was not independently correlated with the OS of patients in the external hospitals (model 2 in Table 5; HR =1.28, 95% CI: 0.86–1.92, P=0.230).

Prognostic significance of CT-defined necrosis

The 1-, 2-, and 3-year DFS rates of patients with PDAC with CT-defined necrosis in the discovery cohort were 49.7%, 21.9%, and 14.0%, respectively, all of which were significantly lower than the corresponding rates of patients without CT-defined necrosis (87.4%, 58.1%, and 29.3%, respectively; P<0.001). In terms of OS, the 1-, 2-, and 3-year DFS rates of patients with PDAC with CT-defined necrosis were 78.1%, 47.5%, and 25.3%, respectively; again, these rates were significantly lower than those of patients without CT-defined necrosis (98.4%, 68.6%, and 43.9%, respectively; P<0.001). The Kaplan-Meier curve also indicated that DFS and OS rates of patients with CT-defined necrosis in our hospital and the external hospitals were significantly lower than those who did not have CT-defined necrosis (both P values <0.001; Figure 4).

Figure 4 Kaplan-Meier estimates of DFS and OS associated with CT-defined necrosis (cutoff point: 15 HU). In both the DFS and OS groups, CT-defined necrosis-negative status was associated with a significantly better prognosis. (A) The DFS of patients in our hospital (n=966; P<0.001, log-rank test). (B) The OS of patients in our hospital (n=966; P<0.001, log-rank test). (C) The DFS of patients in the external hospitals (n=150; P<0.001, log-rank test). (D) The OS of patients in the external hospitals (n=150; P<0.001, log-rank test). (E) The OS of 328 patients with HTN evaluation (P<0.001, log-rank test). (F) The DFS in 328 patients with HTN evaluation (P=0.01, log-rank test). Our hospital: Shengjing Hospital of China Medical University; external hospitals: Guangdong Provincial People’s Hospital and Tianjin Tumor Hospital. CI, confidence interval; CT, computed tomography; DFS, disease-free survival; HR, hazard ratio; HTN, histological tumor necrosis; HU, Hounsfield units; OS, overall survival; PDAC, pancreatic adenocarcinoma.

Discussion

This study investigated the value of CE-CT in evaluating HTN in patients with PDAC and demonstrated that CT-defined necrosis serves as an independent predictor of both DFS and OS in patients with resectable PDAC. We assessed the diagnostic efficacy of different attenuation values of PDAC in the venous phase and unenhanced phase for HTN. Our results showed that the difference in the tumor-enhancement CT value (delta ≤15 HU) indicated significantly better performance for diagnosing HTN than did the other thresholds (AUC: 0.93 vs. 0.75–0.87, P<0.05) and the subjective evaluations of the radiologists (AUC: 0.93 vs. 0.74 and 0.78, P<0.05). CT-defined necrosis was related to poor survival in patients with PDAC after surgery. In the multivariate Cox proportional hazards model, CT-defined necrosis was an independent prognostic factor of DFS in both the preoperative model and comprehensive model (pre- and postoperative) of our hospital and the external hospitals. In our evaluation of the preoperative OS model, we found that CT-defined necrosis was an independent prognostic factor in our hospital and the external hospitals. Finally, in the evaluation of the comprehensive OS model (pre- and postoperative), CT-defined necrosis was an independent influencing factor for patients in our hospital; however, there was no statistical significance in patients in the external hospitals.

Studies on the assessment of pancreatic necrosis according to the degree of CT enhancement have mainly examined pancreatitis and pancreatic cancer. Among the earlier reports, Johnson et al. (24) included 13 patients with acute pancreatitis, and in 7 of these patients, necrosis was detected in the surgical resection specimens, but the corresponding parts of the pancreas on the CT images of these patients revealed no-enhancement areas; moreover, the sensitivity of CT in defining necrosis was 100%. In Bradley et al.’s study (25), all six patients with acute pancreatitis and a pancreatic density of <40 HU in the parenchymal phase had necrosis in the postoperative specimen, with a positive predictive value of 100% for CT values in predicting pancreatic necrosis. More recently, Tsuji et al. (26) created subtraction color map images from unenhanced CT and CE-CT images. The expected upper enhancement threshold (13 HU for pancreatic parenchymal phase images) for pancreatic necrosis was found to correspond to a change in color hue (from blue to purple). Three radiologists were instructed to diagnose pancreatic necrosis when pancreatic parenchyma showed a purple area that was ≥1 cm². Among the 48 patients included in the study, 11 had pancreatic necrosis, and the highest sensitivity, specificity, and accuracy of the three radiologists for predicting pancreatic necrosis using subtractive maps were 100%, 100%, and 100%, respectively. However, Tsuji et al. simply set a threshold based on previous studies, and their method still relied on the subjective recognition of radiologists of color changes to diagnose necrosis, entailing a certain degree of subjectivity.

The enhancement mode of pancreatic cancer can reflect the blood supply of the tumor in vivo. It is widely acknowledged that PDAC is a typical tumor with poor blood supply, which leads to the persistence of hypoxia in the tissue of PDAC and ultimately to the appearance of histological necrosis in the tumor’s center. Theoretically, the necrotic area of the tumor has no blood supply and appears as an unenhanced area on contrast-enhanced images, with a slightly higher density than the cystic component and blurred margins. Previous studies have reported that PEAs can serve as indicators of HTN. Earlier research (15) defined PEAs as areas with obvious low attenuation in enhancement relative to the surrounding tumor; however, this finding relied only on a subjective evaluation by radiologists, without specific reference standards, and PDAC with positivity of PEAs accounted for 25% of 24 cases with PEA-positive PDAC. Both Hattori et al. and Kudo et al. defined PEAs as areas where the tumor shows a significantly lower attenuation as compared with the surrounding tumor tissue in the late phase of CE-CT. Differences in findings across studies are likely due to the inconsistency in the definition of PEAs. Hattori et al. (17) defined PEAs as the areas where tumor enhancement in the late CE-CT phase is <20 HU as compared with that in the unenhanced phase. In their study, they found that HTN-positive PDAC accounted for 33.3% of the 33 PEA-positive PDAC cases. On the other hand, Kudo et al. (27) defined PEAs as the areas where tumor enhancement in the late CE-CT phase is <45 HU as compared with that in the unenhanced phase. They reported that HTN-positive PDAC accounted for 76.5% of 98 cases of PEA-positive PDAC in their study. However, these studies used inconsistent definitions of PEAs (0–45 HU) and varying proportions of HTN-positive PDAC in PEA-positive PDAC (25–76.5%). Moreover, their sample sizes were small (n=33–221). Our study quantitatively examined different delta thresholds for diagnosing HTN and selected the optimal delta threshold through statistical analysis. We found that in cases of delta ≤15 HU, CT-defined necrosis (AUC =0.93) was significantly superior to other delta threshold values (10–30 HU; AUC =0.74–0.88). Furthermore, our study had a larger patient sample size (n=328), which provided better support for the experimental results. We also demonstrated that CT-defined necrosis had better diagnostic efficacy for HTN than did subjective diagnosis by two radiologists (AUC: 0.93 vs. 0.74 and 0.78). This difference in performance might be due to the fact that it is usually difficult for the naked eye to detect attenuation differences <10–15 HU under the setting of the abdominal CT review window (28). Unlike radiologists who measure CT attenuation by placing a region of interest to diagnose pancreatic necrosis, we used 3D-nnUNet to automatically segment PDAC and also used self-written scripts in MATLAB to obtain the tumor’s attenuation difference between the portal venous phase and unenhanced phase. This method effectively avoided errors introduced by the subjective judgments of radiologists.

Previous studies have reported that the degree of enhancement on CE-CT can predict the prognosis of PDAC. Kim et al. (29) found that as compared with the traditional PDAC with weak enhancement, patients with PDAC with equal enhancement compared with the surrounding normal pancreatic tissue had a longer survival time after operation. Fukukura et al. (30) found that patients with PDAC who showed high enhancement in the parenchymal phase had longer survival than did those with low enhancement. However, these findings were based on the subjective assessments by radiologists. Recently, two studies started further optimization and standardization of the prognostic performance of PDAC by using computer-assisted quantitative increments and ratios. Both Koay et al. (31) and Cai et al. (32) examined the CT attenuation differences between the PDAC tumor and the surrounding pancreatic parenchyma on CE-CT and found that patients with larger differences in CT values had lower differentiation, shorter DFS and OS, and worse prognosis. In contrast to the aforementioned studies on enhancement differences between the tumor and normal pancreatic parenchyma, our study focused on the attenuation values of the tumor on CE-CT images during the portal venous and non-enhancement phases. HTN is a well-recognized indicator of poor prognosis in PDAC (12,13), and we found that CT-defined necrosis had a good performance in diagnosing HTN. Therefore, we hypothesized that CT-defined necrosis has a certain guiding significance for predicting the prognosis of patients with PDAC. We found that patients with positive CT-defined necrosis had worse postoperative prognosis and shorter survival time than those diagnosed with no necrosis on CT. Both Kim et al. (33) and Li et al. (34) defined tumor necrosis as an area without enhancement on CE-CT images. Kim et al. (33) found that necrosis was associated with poor tumor differentiation (P<0.001), lymph node metastasis (P=0.01), and lymphovascular invasion (P=0.005) and was also an independent prognostic factor for tumor recurrence-free survival. Li et al. (34) reported that necrosis was a risk factor for recurrence in patients with surgically resectable PDAC. However, these two studies did not assess the correlation between PEAs on CE-CT imaging and HTN. In our study, we investigated the diagnostic performance of different attenuation values between the portal venous and non-enhancement phases on CE-CT images for HTN and selected the optimal attenuation value to improve the performance of CT-defined necrosis in predicting PDAC prognosis.

We found that CT-defined necrosis outperformed histologic necrosis in predicting DFS and OS, which may be attributed to the limitations of pathological sampling and the biological relevance of radiomics phenotypes. No new UNet (nnUNet)-based whole-tumor segmentation was used to determine CT necrosis, whereas 5-mm-thick histological slices were used for determining HTN relied, which risked the omission of focal necrosis; in contrast, whole-tumor 3D evaluation via CT can overcome the sampling bias inherent in pathologic sectioning. Additionally, the CT threshold captures the nonperfused tumor regions pathologically linked to hypoxia-induced aggression, whereas histology identifies only completed necrosis. Two pivotal studies support this conclusion. First, Koay et al. (31) found that histologic sections sampled less than 0.0001% of the entire tumor volume, whereas CT imaging assesses 100% of the tumor. This profound disparity suggests that pathology fails to capture spatially heterogeneous hypoxia—a key driver of treatment resistance. In contrast, CT necrosis quantification enables 3D evaluation of the entire tumor mass, revealing hypoxia gradients that correlate with aggressive biology. Koay et al. further found that among 247 patients with pancreatic cancer examined, CT necrosis quantification provided a significantly higher prognostic accuracy for OS (concordance index =0.79) than did histologic assessment (concordance index =0.62; P<0.001). In a multicenter validation study of 418 surgically resected patients, Cai et al. (32) found that CT-defined necrosis volume >30% was associated with worse DFS (HR =3.21, 95% CI: 2.18–4.73), whereas histologic necrosis showed no association (P=0.37). Notably, CT-defined necrosis exhibited a strong positive correlation with hypoxia-related gene expression signatures (Pearson r=0.71; P<0.001), supporting its value as a noninvasive imaging biomarker of hypoxia-mediated tumor biology.

Our study involved several limitations that should be addressed. First, we employed a retrospective design, which could have introduced bias in the case selection, and thus future prospective studies are needed. Second, we were unable to retrospectively analyze certain PDAC pathological information, and the evaluation of pathological slides of all patients was not possible. Third, neoadjuvant therapy might have an impact on HTN and PEAs, but we excluded this from our analysis. Further studies are needed to explore this potential influence.

Conclusions

We found that CT-defined necrosis can effectively diagnose HTN in patients with PDAC and is an independent prognostic factor for poor prognosis in patients with resectable PDAC. In future studies, the use of CE-CT for evaluating HTN in PDAC might demonstrate clinical significance for assessing patient prognosis and guiding personalized, stratified treatment.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-905/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-905/dss

Funding: This work was supported by the National Natural Science Foundation of China (No. 82471971), the Liaoning Province Science and Technology Joint Plan (Nos. 2023JH2/101700127 and 2023JH2/101700195), the Leading Young Talent Program of Xingliao Yingcai in Liaoning Province (No. XLYC2203037), and the Liaoning Provincial Science and Technology Program (No. 2025-BS-0581).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-905/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study received ethical approval from the Medical Ethics Committee of Shengjing Hospital of China Medical University (ethical approval No. 2021ps830k) and individual consent for this retrospective analysis was waived. All collaborating institutions have received complete study documentation and furnished formal letters of approval.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33.
2. Winter JM, Cameron JL, Campbell KA, Arnold MA, Chang DC, Coleman J, Hodgin MB, Sauter PK, Hruban RH, Riall TS, Schulick RD, Choti MA, Lillemoe KD, Yeo CJ. 1423 pancreaticoduodenectomies for pancreatic cancer: A single-institution experience. J Gastrointest Surg 2006;10:1199-210; discussion 1210-1.
3. Groot VP, Rezaee N, Wu W, Cameron JL, Fishman EK, Hruban RH, Weiss MJ, Zheng L, Wolfgang CL, He J. Patterns, Timing, and Predictors of Recurrence Following Pancreatectomy for Pancreatic Ductal Adenocarcinoma. Ann Surg 2018;267:936-45.
4. Conroy T, Hammel P, Hebbar M, Ben Abdelghani M, Wei AC, Raoul JL, et al. FOLFIRINOX or Gemcitabine as Adjuvant Therapy for Pancreatic Cancer. N Engl J Med 2018;379:2395-406.
5. van Roessel S, Kasumova GG, Verheij J, Najarian RM, Maggino L, de Pastena M, et al. International Validation of the Eighth Edition of the American Joint Committee on Cancer (AJCC) TNM Staging System in Patients With Resected Pancreatic Cancer. JAMA Surg 2018;153:e183617.
6. Kamarajah SK, Burns WR, Frankel TL, Cho CS, Nathan H. Validation of the American Joint Commission on Cancer (AJCC) 8th Edition Staging System for Patients with Pancreatic Adenocarcinoma: A Surveillance, Epidemiology and End Results (SEER) Analysis. Ann Surg Oncol 2017;24:2023-30.
7. Mitsunaga S, Hasebe T, Kinoshita T, Konishi M, Takahashi S, Gotohda N, Nakagohri T, Ochiai A. Detail histologic analysis of nerve plexus invasion in invasive ductal carcinoma of the pancreas and its prognostic impact. Am J Surg Pathol 2007;31:1636-44.
8. Lüttges J, Schemm S, Vogel I, Hedderich J, Kremer B, Klöppel G. The grade of pancreatic ductal carcinoma is an independent prognostic factor and is superior to the immunohistochemical assessment of proliferation. J Pathol 2000;191:154-61.
9. Adsay NV, Basturk O, Bonnett M, Kilinc N, Andea AA, Feng J, Che M, Aulicino MR, Levi E, Cheng JD. A proposal for a new and more practical grading scheme for pancreatic ductal adenocarcinoma. Am J Surg Pathol 2005;29:724-33.
10. Takai S, Satoi S, Toyokawa H, Yanagimoto H, Sugimoto N, Tsuji K, Araki H, Matsui Y, Imamura A, Kwon AH, Kamiyama Y. Clinicopathologic evaluation after resection for ductal adenocarcinoma of the pancreas: a retrospective, single-institution experience. Pancreas 2003;26:243-9.
11. Lim JE, Chien MW, Earle CC. Prognostic factors following curative resection for pancreatic adenocarcinoma: a population-based, linked database analysis of 396 patients. Ann Surg 2003;237:74-85.
12. Hiraoka N, Ino Y, Sekine S, Tsuda H, Shimada K, Kosuge T, Zavada J, Yoshida M, Yamada K, Koyama T, Kanai Y. Tumour necrosis is a postoperative prognostic marker for pancreatic cancer patients with a high interobserver reproducibility in histological evaluation. Br J Cancer 2010;103:1057-65.
13. Mitsunaga S, Hasebe T, Iwasaki M, Kinoshita T, Ochiai A, Shimizu N. Important prognostic histological parameters for patients with invasive ductal carcinoma of the pancreas. Cancer Sci 2005;96:858-65.
14. Tempero MA. NCCN Guidelines Updates: Pancreatic Cancer. J Natl Compr Canc Netw 2019;17:603-5.
15. Demachi H, Matsui O, Kobayashi S, Akakura Y, Konishi K, Tsuji M, Miwa A, Miyata S. Histological influence on contrast-enhanced CT of pancreatic ductal adenocarcinoma. J Comput Assist Tomogr 1997;21:980-5.
16. Furukawa H, Takayasu K, Mukai K, Kanai Y, Inoue K, Kosuge T, Ushio K. Late contrast-enhanced CT for small pancreatic carcinoma: delayed enhanced area on CT with histopathological correlation. Hepatogastroenterology 1996;43:1230-7.
17. Hattori Y, Gabata T, Zen Y, Mochizuki K, Kitagawa H, Matsui O. Poorly enhanced areas of pancreatic adenocarcinomas on late-phase dynamic computed tomography: comparison with pathological findings. Pancreas 2010;39:1263-70.
18. Tempero MA, Malafa MP, Al-Hawary M, Behrman SW, Benson AB, Cardin DB, et al. Pancreatic Adenocarcinoma, Version 2.2021, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2021;19:439-57.
19. Yao J, Cao K, Hou Y, Zhou J, Xia Y, Nogues I, et al. Deep Learning for Fully Automated Prediction of Overall Survival in Patients Undergoing Resection for Pancreatic Cancer: A Retrospective Multicenter Study. Ann Surg 2023;278:e68-79.
20. Zhao R, Jia Z, Chen X, Ren S, Cui W, Zhao DL, Wang S, Wang J, Li T, Zhu Y, Tang X, Wang Z. CT and MR imaging features of pancreatic adenosquamous carcinoma and their correlation with prognosis. Abdom Radiol (NY) 2019;44:2822-34.
21. Ding Y, Zhou J, Sun H, He D, Zeng M, Rao S. Contrast-enhanced multiphasic CT and MRI findings of adenosquamous carcinoma of the pancreas. Clin Imaging 2013;37:1054-60.
22. Hori S, Shimada K, Ino Y, Oguro S, Esaki M, Nara S, Kishi Y, Kosuge T, Hattori Y, Sukeda A, Kitagawa Y, Kanai Y, Hiraoka N. Macroscopic features predict outcome in patients with pancreatic ductal adenocarcinoma. Virchows Arch 2016;469:621-34.
23. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45.
24. Johnson CD, Stephens DH, Sarr MG. CT of acute pancreatitis: correlation between lack of contrast enhancement and pancreatic necrosis. AJR Am J Roentgenol 1991;156:93-5.
25. Bradley EL 3rd, Murphy F, Ferguson C. Prediction of pancreatic necrosis by dynamic pancreatography. Ann Surg 1989;210:495-503; discussion 503-4.
26. Tsuji Y, Takahashi N, Fletcher JG, Hough DM, McMenomy BP, Lewis DM, Vege SS, Chari ST, McCollough CH, Grant KL, Klotz E. Subtraction color map of contrast-enhanced and unenhanced CT for the prediction of pancreatic necrosis in early stage of acute pancreatitis. AJR Am J Roentgenol 2014;202:W349-56.
27. Kudo M, Kobayashi T, Gotohda N, Konishi M, Takahashi S, Kobayashi S, Sugimoto M, Okubo S, Martin J, Cabral H, Ishii G, Kojima M. Clinical Utility of Histological and Radiological Evaluations of Tumor Necrosis for Predicting Prognosis in Pancreatic Cancer. Pancreas 2020;49:634-41.
28. Baron RL. Understanding and optimizing use of contrast material for CT of the liver. AJR Am J Roentgenol 1994;163:323-31.
29. Kim JH, Park SH, Yu ES, Kim MH, Kim J, Byun JH, Lee SS, Hwang HJ, Hwang JY, Lee SS, Lee MG. Visually isoattenuating pancreatic adenocarcinoma at dynamic-enhanced CT: frequency, clinical and pathologic characteristics, and diagnosis at imaging examinations. Radiology 2010;257:87-96.
30. Fukukura Y, Takumi K, Higashi M, Shinchi H, Kamimura K, Yoneyama T, Tateyama A. Contrast-enhanced CT and diffusion-weighted MR imaging: performance as a prognostic factor in patients with pancreatic ductal adenocarcinoma. Eur J Radiol 2014;83:612-9.
31. Koay EJ, Lee Y, Cristini V, Lowengrub JS, Kang Y, Lucas FAS, et al. A Visually Apparent and Quantifiable CT Imaging Feature Identifies Biophysical Subtypes of Pancreatic Ductal Adenocarcinoma. Clin Cancer Res 2018;24:5883-94.
32. Cai X, Gao F, Qi Y, Lan G, Zhang X, Ji R, Xu Y, Liu C, Shi Y. Pancreatic adenocarcinoma: quantitative CT features are correlated with fibrous stromal fraction and help predict outcome after resection. Eur Radiol 2020;30:5158-69.
33. Kim DW, Lee SS, Kim SO, Kim JH, Kim HJ, Byun JH, Yoo C, Kim KP, Song KB, Kim SC. Estimating Recurrence after Upfront Surgery in Patients with Resectable Pancreatic Ductal Adenocarcinoma by Using Pancreatic CT: Development and Validation of a Risk Score. Radiology 2020;296:541-51.
34. Li D, Wang L, Cai W, Liang M, Ma X, Zhao X. Prognostic stratification in patients with pancreatic ductal adenocarcinoma after curative resection based on preoperative pancreatic contrast-enhanced CT findings. Eur J Radiol 2022;151:110313.