Computed tomography as a method for diagnosing true vascular invasion in pancreatic cancer

Full Text

BACKGROUND

Pancreatic cancer is one of the most aggressive malignancies, characterized by a five-year survival rate of less than 10% and high mortality even with multimodal treatment. Prognostic outcome is critically determined by the feasibility of achieving a radical (R0) resection, which is directly contingent upon the presence or absence of histologically confirmed true vascular invasion [1]. Involvement of the portal venous system, as assessed by preoperative imaging modalities, along with arterial involvement, forms the basis of the current classification of pancreatic cancer resectability (resectable, borderline resectable, locally advanced). However, unlike arterial invasion, tumor extension within the lumen of blood vessels is not considered a contraindication to surgical treatment [2].

Multidetector computed tomography (MDCT) is currently considered the gold standard for assessing venous invasion in the initial staging of pancreatic cancer and for determining treatment strategy [3]. Literature reviews and clinical evidence indicate that multidetector computed tomography (MDCT) exhibits high diagnostic accuracy for evaluating great vessel invasion, with reported specificity ranging from 82% to 100% and sensitivity from 70% to 96% [4]. Nevertheless, there is a significant percentage of discrepancies between radiological assessment and morphologically confirmed true venous wall invasion [5]. Classical CT criteria (tumor-vessel contact area, vessel deformation and lumen narrowing, occlusion, irregular margins) only indirectly correlate with morphological signs of invasion [6], leading to both overdiagnosis (incorrect reclassification of patients as borderline resectable or unresectable) and underdiagnosis, with surgical interventions performed without a realistic prospect of achieving an R0 resection [7].

In recent years, neoadjuvant polychemotherapy (PCT) has been increasingly used in patients with borderline resectable and locally advanced pancreatic cancer, and in some cases—in patients with initially resectable tumors [8]. This approach allows for an increased rate of R0 resections, a reduced risk of early progression, and identification of patients with unfavorable tumor biological characteristics prior to surgery [8]. As a result, an increasing number of patients undergo surgery after courses of systemic treatment, which fundamentally changes the conditions of preoperative diagnosis.

Following neoadjuvant PCT, pronounced fibrous and inflammatory changes develop in the tumor area and adjacent vessels, accompanied by remodeling of the parapancreatic tissue and vascular wall. These processes can significantly distort CT imaging findings, making it difficult to differentiate between tumor tissue and post-treatment fibrosis. Consequently, conventional CT-based criteria used to evaluate vascular invasion, which were established primarily in chemotherapy-naive patient populations, may exhibit reduced diagnostic accuracy in patients who have undergone neoadjuvant therapy [9]. Given these limitations, a systematic assessment of preoperative multidetector computed tomography (MSCT) for detecting true portal vein system invasion in pancreatic cancer patients after neoadjuvant polychemotherapy (PCT) is essential. This evaluation must incorporate mandatory correlation of imaging findings with histological results [10]. These findings have the potential to enhance clinical decision-making regarding treatment strategies, refine indications for venous resection procedures, and decrease the incidence of unwarranted refusals to perform potentially curative surgery.

Aim: The primary aim of this study is to evaluate the impact of neoadjuvant polychemotherapy (PCT) on the diagnostic accuracy of preoperative contrast-enhanced multidetector computed tomography (MSCT) for detecting true venous invasion in patients with pancreatic cancer. This will be accomplished by systematically analyzing imaging features and correlating them with histological findings from resected venous segments.

METHODS

Study Design

This study represents a retrospective case series analysis of 61 patients with ductal adenocarcinoma of the pancreatic head who demonstrated CT-confirmed tumor contact with a main venous structure of the mesentericoportal system. Patients were allocated to two groups according to the selected treatment strategy. Group 1 comprised 37 patients in whom surgery was performed as the primary intervention. Group 2 comprised 24 patients who received neoadjuvant polychemotherapy (PCT) according to the FOLFIRINOX regimen (2 to 8 cycles) prior to surgical resection.

Study Setting

All medical care was provided at the Research Institute of Clinical Oncology “Nizhny Novgorod Regional Clinical Oncological Dispensary”, Nizhny Novgorod, Russia.

Eligibility Criteria

Inclusion criteria: age above 18 years; histologically verified ductal adenocarcinoma of the pancreas; performance of pancreaticoduodenectomy with resection of the mesentericoportal venous segment followed by histopathological examination of the resected specimen; performance of contrast-enhanced multidetector computed tomography (MDCT) within 4 weeks prior to surgery; performance of contrast-enhanced MDCT prior to initiation of treatment in patients who received neoadjuvant PCT.

There were no criteria for exclusion.

Intervention

All patients underwent pancreaticoduodenectomy with excision of the main venous wall, followed by histopathological evaluation of the resected vessel for the presence of true tumor invasion. In Group 1, MDCT was performed preoperatively within 2 weeks of surgery. In Group 2, MDCT was performed twice: within 2 weeks before initiation of treatment and within 4 weeks prior to surgery. In accordance with National Comprehensive Cancer Network (NCCN) guidelines (version 2.2025), indications for surgery following neoadjuvant PCT included the absence of overt systemic disease progression and at least stable or declining levels of the tumor marker CA 19-9 (Cancer Antigen 19-9).

Study Outcomes

During CT image analysis, the axial length of tumor-vessel contact, the degree of vascular encasement, and the presence of venous wall contour irregularity were assessed. All CT series were independently reviewed by two expert radiologists with more than 10 years of relevant experience. Following histopathological examination, both groups were subdivided according to the presence or absence of true vascular invasion (morphologically confirmed tumor growth into the vascular wall).

Statistical Analysis

Planned sample size. Sample size was not calculated prospectively.

Statistical Methods. Categorical data are presented as absolute frequencies (n) and relative proportions (%). Comparative analysis of categorical variables was performed using the Chi-square test (χ²) and Fisher’s exact test. Binary logistic regression without regularization and with a probability threshold of 0.5 was applied. Statistical significance was defined as p <0.05. All analyses were conducted using the open-source statistical computing environment R in RStudio IDE, version 2024.04.2+764 (Posit Software, PBC).

RESULTS

Study Population Characteristics

A retrospective analysis was performed on data from 61 patients with pancreatic head cancer following radical surgical treatment. Patients were divided into two groups: those without preoperative systemic therapy (n=37) and those who received neoadjuvant polychemotherapy (n=24). The mean age of study participants was 62.8±10.3 years; 29 patients (47.5%) were male.

Primary Results

Up-front surgery group (Group 1). In the first stage, we validated established radiological criteria for venous invasion of pancreatic cancer in 37 patients who underwent surgery without prior systemic treatment. Patients were stratified into two subgroups based on morphologically confirmed true vascular invasion. On preoperative MDCT, tumor contact with the portal venous system was identified in all patients in this group. True venous invasion was confirmed histopathologically in 23/37 (62.2%) cases, whereas in 14 (37.8%) patients, tumor growth into the vessel wall was not detected (Table 1).

 

Table 1

Assessment of tumor–major venous vessel interface in patients undergoing upfront surgical resection without prior neoadjuvant chemotherapy

CT features	Presence of vascular invasion
	True vascular wall invasion, n (%)		p-value (χ2)	Odds Ratio (logistic regression)
	Present (n=23)	Absent (n=14)
Extent of involvement >180°	7 (30.4)	0	0.022	3.629
Extent of lesion >2 cm	13 (56.5)	4 (28.6)	0.099	0.453
Presence of vessel wall erosion	14 (60.9)	0	<0.001	>108
Tumor contact with the superior mesenteric vein	12 (52.2)	3 (21.4)	0.065	5.483
Tumor contact with the portal vein	6 (26.1)	6 (42.9)	0.291	-
Tumor contact with the superior mesenteric-portal vein confluence	5 (21.7)	5 (35.7)	0.354	-

Note. CT, computed tomography.

 

Radiological analysis identified focal disruption of the venous wall contour and tumor encasement of more than 180° around the vessel as independent imaging markers of true vascular invasion. However, the combination of a smooth venous contour with encasement of less than a semicircle did not exclude transmural tumor growth: in our cohort of 23 such patients, invasion was histologically confirmed in 9 (39.1%).

Notably, an axial contact length exceeding 2 cm did not reach statistical significance, while in the setting of a smooth contour and encasement below 180°, it was indicative of the absence of tumor ingrowth. A relationship between the frequency of invasion and the anatomical segment of the mesentericoportal trunk was also identified: the highest incidence was observed with superior mesenteric vein involvement (80%), whereas for the portal vein and the confluence zone the rate was 50% in each case.

Binary logistic regression analysis yielded the following predictive model:

P (invasion)=1/(1+exp(−(−0.8775+1.2889×

(encasement >180°)−0.7917×

(contact length >2 cm)+25.0090×

(vascular wall erosion)+1.7017×

(tumor contact with superior mesenteric vein)))).

Model accuracy was 81.1%, sensitivity 82.6%, and specificity 78.6%. Overall statistical significance of the model: p=0.00024. A ROC curve was constructed: AUC=0.888 (Fig. 1).

 

Fig. 1. ROC curve showing the relationship between the presence of true venous invasion and radiological features in pancreatic head cancer.

 

In the second stage of the study, we assessed radiological features of venous invasion following neoadjuvant PCT in 24 patients. As in Group 1, patients were stratified into two subgroups based on the presence of true vascular invasion (Table 2). The mean number of neoadjuvant FOLFIRINOX cycles in the subgroup with venous involvement was 4 (range 2–6), compared with 6 (range 6–8) in patients without evidence of vascular invasion. Administration of more than 5 cycles of neoadjuvant PCT was a statistically significant predictor of the absence of true invasion (p <0.05).

 

Table 2

Characteristics of tumor contact with a major venous vessel in patients who underwent neoadjuvant chemotherapy

Detection of vascular invasion prior to/following/ during neoadjuvant polychemotherapy		True vascular wall invasion, n (%)		p-value
		Present (n=8)	Absent (n=16)
Extent of involvement >180°	prior to	5 (62.5)	8 (50)	0.563
	following	4 (50)	7 (43.75)	0.773
Reduction in volume	during neoadjuvant polychemotherapy	1 (12.5)	1 (6.25)	0.602
Length of involvement >2 cm	prior to	7 (87.5)	14 (43.75)	1.000
	following	5 (62.5)	13 (81.25)	0.318
Decrease in length <2 cm	during neoadjuvant polychemotherapy	2 (25)	1 (6.25)	0.191
Detection of vascular wall ulceration	prior to	6 (75)	12 (75)	1.000
	following	2 (25)	11 (68.75)	0.043
Contour alignment	during neoadjuvant polychemotherapy	4 (50)	1 (6.25)	0.013
Response to neoadjuvant chemotherapy (according to RECIST 1.1)		3 (37.5)	8 (50)	0.563
Administration of >5 cycles of neoadjuvant chemotherapy		2 (25)	16 (100)	<0.05

 

Based on preoperative radiological evaluation, 13 (54.2%) patients demonstrated stable disease in response to neoadjuvant PCT per Response Evaluation Criteria in Solid Tumors (RECIST 1.1), of whom true vascular invasion was confirmed in 5 and not confirmed in 7. Partial response was observed in 11 (45.8%) patients: true vascular invasion was confirmed in 3 and not confirmed in 8, which did not reach statistical significance (p >0.05).

During the course of neoadjuvant PCT, a reduction in the degree of venous encasement by the tumor was noted in 2 cases (postoperatively, one with morphologically confirmed invasion and one without), a reduction in contact length to below 2 cm was observed in 3 cases (1 without and 2 with invasion), and smoothing of the venous contour occurred in 5 patients (1 without and 4 with invasion). Nevertheless, analysis of all MDCT series (pre- and post-neoadjuvant PCT) in patients who received neoadjuvant PCT demonstrated that none of the investigated radiological criteria reliably predicted the presence or absence of true vascular invasion. Predictors of morphologically confirmed invasion previously validated in chemotherapy-naive patients—namely, venous wall erosion and tumor encasement exceeding 180°—did not emerge as significant markers of true vascular invasion in the post-treatment setting. Binary logistic regression analysis revealed that a predictive model based on these features was statistically non-significant (p=0.61). This situation is well illustrated by a clinical example (Fig. 2), in which the presence of all absolute radiological features of venous invasion was not corroborated by histopathology.

 

Fig. 2. Extended tumor contact with the superior mesenteric vein and vessel wall ulceration (arrow) on preoperative multidetector computed tomography (MDCT), in the absence of true invasion according to postoperative morphological examination in a patient who received 6 cycles of neoadjuvant polychemotherapy.

 

Paradoxically, a model incorporating favorable radiological features—specifically, contour smoothing, reduction in contact length, and reduction in the degree of circumferential encasement—demonstrated some predictive value (p=0.04; accuracy 79.2%, sensitivity 50%, specificity 93.8%) with respect to the absence of true transmural growth.

DISCUSSION

Summary оf Principal Findings

In pancreatic head cancer, preoperative chemotherapy according to the FOLFIRINOX regimen substantially undermines the ability to determine true vascular invasion based on preoperative MDCT findings. The presence of venous wall erosion and circumferential encasement exceeding 180° in these cases no longer serves as a reliable indicator of true vessel wall involvement and may only assist in planning the extent of vascular reconstruction. Administration of more than 5 cycles of neoadjuvant PCT is associated with a significantly reduced probability of histologically confirmed tumor invasion of the venous wall.

Interpretation

Despite the inherent limitations of this study, the findings indicate that computed tomography is not an adequate method for evaluating local tumor status in the context of neoadjuvant PCT.

The present study demonstrates that the diagnostic value of preoperative computed tomography in detecting true venous invasion in pancreatic cancer patients is substantially reduced following neoadjuvant PCT. The principal finding is the absence of a statistically significant association between classical CT features of vascular involvement (the angle of circumferential tumor contact, axial contact length, and character of the venous wall contour) and morphologically confirmed tumor growth into the vessel wall in the post-neoadjuvant group [11, 12].

This finding has important clinical implications, given that these very features are traditionally employed for resectability assessment and surgical decision-making. The current data indicate that, following neoadjuvant PCT, the persistence of suspicious CT features cannot be regarded as a reliable indicator of true venous invasion, and their absence cannot be interpreted as proof of vascular wall integrity. Our findings confirm that standard radiological criteria developed predominantly for chemotherapy-naïve patients cannot be directly extrapolated to patients who have undergone preoperative systemic therapy [13, 14].

The inverse relationship identified in our study between the number of neoadjuvant PCT cycles and the incidence of morphologically confirmed venous invasion may reflect both the biological effect of systemic therapy and a patient selection phenomenon. More prolonged treatment likely contributes to reduction of the neoplastic component within the zone of vascular contact and to the predominance of fibrotic-inflammatory changes, thereby reducing the incidence of true tumor ingrowth into the vessel wall. At the same time, the wide confidence intervals and the retrospective nature of the analysis preclude drawing definitive conclusions, and further confirmation in prospective studies is required.

The clinical example presented in Fig. 2 vividly illustrates the problem of venous invasion overdiagnosis following neoadjuvant therapy. Extensive and protracted tumor contact with the superior mesenteric vein accompanied by wall erosion on CT—in the absence of histologically confirmed tumor growth— highlights the limitations of visual assessment in the context of post-therapeutic fibrosis. Such scenarios carry significant practical consequences, as they may lead to unjustified refusal of surgical treatment or an excessively pessimistic assessment of resectability.

In the comparative group of patients without preoperative systemic treatment, classical CT features—including irregularity of the venous wall contour and tumor encasement exceeding 180°— demonstrated a pronounced association with true venous invasion. Nevertheless, even in this cohort, the absence of these findings did not reliably exclude tumor growth into the vascular wall, which is consistent with the published literature describing high specificity but limited sensitivity of MDCT for diagnosing venous involvement. This underscores that the discordance between radiological and morphological assessment of vascular invasion is not an exclusive consequence of neoadjuvant therapy, yet is substantially more pronounced in the post-neoadjuvant group [15]. Although MDCT represents the gold standard for diagnosing pancreatic cancer and is utilized to monitor treatment response, the present findings underscore the need to reassess the interpretation of preoperative MDCT in patients following neoadjuvant PCT. In such clinical scenarios, MDCT should be regarded not as a definitive tool for resectability determination, but rather as a means of surgical planning. In the absence of evidence of systemic progression, a strategy of active surgical intervention appears more justified—even in the presence of adverse CT features of vascular contact [16].

CONCLUSION

The results of the present study demonstrate that, following neoadjuvant polychemotherapy, computed tomography has limited capabilities in diagnosing true venous invasion in pancreatic cancer. This consideration must be incorporated into resectability stratification and surgical planning in order to avoid unwarranted refusals of potentially radical interventions and to ensure an optimal balance between oncological radicality and surgical feasibility.

Additional information

Author contributions: D.M. Kuchin, data and evidence collection, analysis and interpretation of the obtained data, participation in the scientific design, formulation and development of the key goals and objectives, writing the article; G.A. Yarikhovich, literature search and analysis, interpretation of the research results, their processing, writing and editing the text; A.V. Smirnov, participation in the analysis and interpretation of the data, editing the text; Ya.I. Kolesnik, data and evidence collection, interpretation of the research results; V.E. Zagaynov, concept development, formulation and development of the key goals and objectives, critical revision with the introduction of valuable comments on the content, accepting responsibility for all aspects of the work, the integrity of all parts of the article, and its final version. All authors provided approval of the version to be published and agree to be 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.

Ethics approval: The study was approved by the Local Ethics Committee of Nizhny Novgorod Regional Clinical Oncological Dispensary (Extract No. 12 from Protocol No. 21 dated January 19, 2023). All the research participants have voluntarily signed the informed consent form prior to the inclusion into the research program.

Funding source: The research and publication of this article were financed by the authors themselves.

Disclosure of interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Statement of originality: The authors did not utilize previously published information (text, illustrations, data) in conducting the research and creating this paper.

Data availability statement: The editorial policy regarding data sharing does not apply to this work, data can be published as open access.

Generative AI: Generative AI technologies were not used for this article creation.

About the authors

Denis M. Kuchin

Research Institute of Clinical Oncology «Nizhny Novgorod Regional Clinical Oncological Dispensary»; Privolzhsky Research Medical University; Nizhny Novgorod Regional Clinical Hospital named after N.A. Semashko

Author for correspondence.
Email: pomc.kuchin@gmail.com
ORCID iD: 0000-0002-4148-2953
SPIN-code: 6815-9959

MD, PhD

Russian Federation, Nizhny Novgorod; Nizhny Novgorod; Nizhny Novgorod

Grigory A. Yarikhovich

Research Institute of Clinical Oncology «Nizhny Novgorod Regional Clinical Oncological Dispensary»; Privolzhsky Research Medical University

Email: yandr24@gmail.com
ORCID iD: 0000-0002-6824-6379
SPIN-code: 2806-8721
Russian Federation, Nizhny Novgorod; Nizhny Novgorod

Alexander V. Smirnov

Federal Research and Clinical Center of Specialized Medical Care and Medical Technologies

Email: smirnov.av@fnkc-fmba.ru
ORCID iD: 0000-0003-3897-8306
SPIN-code: 5619-1151

MD, PhD, Assistant Professor

Russian Federation, Moscow

Yan I. Kolesnik

Research Institute of Clinical Oncology «Nizhny Novgorod Regional Clinical Oncological Dispensary»; Privolzhsky Research Medical University

Email: kolesnik-y-i@yandex.ru
ORCID iD: 0000-0002-7959-1813
SPIN-code: 9540-8042

MD, PhD

Russian Federation, Nizhny Novgorod; Nizhny Novgorod

Vladimir E. Zagainov

Research Institute of Clinical Oncology «Nizhny Novgorod Regional Clinical Oncological Dispensary»; Privolzhsky Research Medical University

Email: zagainov@xmail.ru
ORCID iD: 0000-0002-5769-0378
SPIN-code: 6477-0291

MD, PhD, Assistant Professor

Russian Federation, Nizhny Novgorod; Nizhny Novgorod

References

Supplementary files