Volume and functional changes of remnant pancreas after different types of pancreatectomy: Exploring the regenerative potential

Abstract

Evidence on pancreatic regeneration and functional recovery after pancreatectomy remains limited. This study investigates the correlation between volumetric changes and endocrine function of the remnant pancreas following pancreaticoduodenectomy (PD) and distal pancreatectomy (DP). We conducted a retrospective cohort study of patients who underwent PD or DP between January 2009 and December 2017 at a single institution. Pancreatic volume was assessed using contrast-enhanced computed tomography preoperatively and at 3 months, 1 year, and 2 years postoperatively. Endocrine function was evaluated using serial C-peptide index (CPI) measurements. A total of 90 PD and 45 DP patients were analyzed. Compared with initial residual volume, remnant pancreas in PD progressively declined (80.79%, 68.67%, and 65.34% at 3 months, 1 year, and 2 years; β = −11.85, p < 0.001), whereas DP patients demonstrated hypertrophic changes (106.25%, 106.62%, and 110.43%; β = 2.97, p = 0.019). New-onset diabetes occurred in 22.7% of PD and 33.3% of DP patients. Postoperative pancreatic duct dilatation was associated with greater atrophy in PD (β = −9.82, p = 0.027). Despite superior volume preservation in DP, PD demonstrated better endocrine functional recovery (CPI/volume ratio: β = 126.9, p < 0.001), corresponding with lower new-onset diabetes incidence. Pancreatic volume and endocrine function recover independently after pancreatectomy. Despite greater volume preservation, DP patients experience more endocrine dysfunction than PD due to higher islet density in the distal pancreas.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36886-4.

Introduction

Pancreatectomy is a standard treatment for several indications, such as pancreatic adenocarcinoma, pancreatic adenoma, pancreatic cysts, and chronic pancreatitis. There are two primary surgical methods of pancreatectomy, pancreaticoduodenectomy (PD) and distal pancreatectomy (DP), depending on the location and extent of the pancreatic lesion. Pancreatectomy is much safer nowadays; the mortality rate of PD has even decreased to 0% ~ 5% in high volume medical centers¹. The improvement of survival rate after pancreatectomy had raised the consideration of long-term metabolic consequences.

Pancreas is an important digestive organ comprising two major components: endocrine glands to regulate metabolism and exocrine glands to secrete digestive enzymes². Islets of Langerhans is a cluster of endocrine cell, only account for 1–2% of pancreas parenchyma but play a critical role in glucose regulation³. Furthermore, β cells make up 50 ~ 70% of the cells in pancreatic islets that synthesize and secrete insulin to control glucose metabolism². The loss of pancreatic parenchyma after pancreatectomy would lead to deterioration of glucose homeostasis due to decrease of β cells amount. Some studies have shown that glucose levels start to rise when β cells mass is reduced more than 50% compared with normal^4,5. About 5–10% of diabetes mellitus cases are related to pancreatic diseases in the Western country⁶. Those new-onset diabetes mellitus after pancreatectomy is now defined as pancreatogenic diabetes mellitus (Type 3c DM)⁶.

Unlike the well-documented regenerative capability of the liver after major hepatectomy, the capacity for pancreatic regeneration following pancreatectomy remains controversial. This raises the question of whether changes in pancreatic volume after pancreatectomy also correlate with functional recovery of the pancreas. In mice and rats, 90% pancreatectomy can induce ductal cell proliferation and upregulate embryonic gene expression, such as Pdx1, Ngn3, and Ptf1a in mature duct cells, to promote the regenerative process⁷. However, the regenerative capability of the pancreas is decreased with age and limited in adult rats⁸. Similarly, previous observation from children suggested tissue regeneration after near-total pancreatectomy⁹, but the capacity for pancreatic regeneration is absent in adult humans¹⁰. You et al. reported that the remnant pancreas atrophied significantly at long-term follow-up after PD¹¹. However, until now, there is no study on volume change of the remnant pancreas after DP. Embryonically, the head pancreas and distal pancreas (neck, body, and tail) are derived from the different origin¹². Hence, we conjecture the different parts of the pancreas may display different regenerative capabilities. Although the deterioration of glucose homeostasis after pancreatectomy had been demonstrated^4,5, restoration of endocrine functions after different procedures remains unclear. Therefore, in this study, we tried to compare the long-term changes in pancreatic volume and endocrine function between PD and DP, and to investigate whether postoperative volume changes correlate with endocrine function recovery and explore the regenerative potential.

Materials and methods

Study design and participants

This was a single-center retrospective observational cohort study conducted at National Cheng Kung University Hospital. Patients who underwent pancreaticoduodenectomy (PD) or distal pancreatectomy (DP) between January 2009 and December 2017 were identified from institutional surgical and radiologic databases. The study protocol was reviewed and approved by the Institutional Review Board of National Cheng Kung University Hospital (IRB No. B-ER-108–312), and the requirement for informed consent was waived due to the retrospective nature of the study. All methods were performed in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki and institutional policies for research involving human subjects.

To evaluate longitudinal volumetric and endocrine changes, only patients with complete preoperative data and serial postoperative follow-up, including abdominal CT volumetry and metabolic assessments (C-peptide, fasting glucose, HbA1c) for at least 1 year were included. For malignant disease, a minimum of 2-year disease-free survival was required to minimize confounding from tumor recurrence on pancreatic morphology and glucose metabolism.

Clinical variables collected included age, sex, body mass index (BMI), pathological diagnosis (benign vs. malignant), surgical procedure, neoadjuvant therapy, preoperative metabolic status, and perioperative conditions including pancreatitis, and obstructive jaundice. In patients with DP, outcomes were additionally compared between those with and without spleen preservation. Standard demographic and laboratory information was retrieved from electronic medical records. Diagnosis of diabetes and metabolic outcomes were defined according to the American Diabetes Association (ADA) criteria. New-onset diabetes was diagnosed in previously non-diabetic patients when fasting glucose ≥ 126 mg/dL, HbA1c ≥ 6.5%, random glucose ≥ 200 mg/dL with symptoms, or initiation of glucose-lowering medication was documented, with confirmation on a separate test where applicable. Diabetes remission was defined as HbA1c < 6.5% in the absence of antidiabetic medication for ≥ 3 months¹³. Pre-operative duct size was recorded as baseline anatomy; post-operative pancreatic duct dilatation > 3 mm on CT/MRI during follow-up was regarded as radiologic evidence of PJ stricture, consistent with prior literature-based criteria^14–16.

Follow-up abdominal computed tomography (CT) was typically performed at 3, 12, and 24 months after surgery for benign disease, and at 3, 6, 9, 12, 18, and 24 months for malignant disease.

Outcomes

The primary outcome was volumetric change of the remnant pancreas, evaluated by CT at baseline (preoperative) and at 3 months, 1 year, and 2 years postoperatively. Secondary outcomes included postoperative endocrine function, assessed by serial C-peptide, fasting glucose, HbA1c, and C-peptide index (CPI) measurements at corresponding time points. CPI was defined as C-peptide (ng/ml) divided by blood glucose (mg/dL) and multiplied by 100¹⁶. The relationships between pancreatic volume, glycemic status, and endocrine function were subsequently analyzed.

Image/CT volumetry

Contrast-enhanced abdominal CT was used for quantitative volumetry using INFINITT PACS medical image processing software (INFINITT Healthcare Co., Ltd., Korea). Preoperative whole-pancreas volume and postoperative remnant volume at 3 months, 1 year, and 2 years were measured on axial images reconstructed at 3-mm slice thickness. In each slice, the pancreas parenchyma was manually traced, excluding portal/splenic vessels and tumor tissue. In malignant lesions, which often present hypodense and infiltrative margins, the segmentation boundary was drawn along the most solid and clearly defined contour to avoid inclusion of peri-tumoral infiltration. The main pancreatic duct was generally negligible in non-dilated glands (1–2 mm) and therefore not specifically removed; however, when duct dilatation exceeded 3 mm, the duct lumen was excluded from the parenchymal mask to avoid overestimation of residual pancreatic tissue. The pancreatic volume of each slice was calculated as pancreatic area (mm²) × slice thickness (3 mm), and total volume was obtained by summing all segmented slices.

Immediate postoperative CT was not routinely performed in our institution unless complications were suspected to avoid unnecessary radiation exposure and contrast-related risks. Early postoperative scans may also contain edema, inflammatory infiltration, and fluid collections that obscure anatomical margins and lead to overestimation of pancreatic volume. Therefore, the pancreatic transection margin was first identified on the earliest postoperative CT obtained at 1–3 months, and then retro-matched to the corresponding anatomical level on the preoperative CT to estimate the initial residual pancreatic volume (IRV = 100%), which served as the standardized baseline for longitudinal comparison. Subsequent postoperative volumetry represented proportional change from IRV, reported as residual volume (V%) and postoperative volume-change ratio (V-change%).

Figure 1 illustrates the volumetry workflow. Panels A and C show preoperative segmentation before PD and DP, with the tumor (T), portal vein (PV), splenic vein (SV), and resection margin (red arrow) identified. Panels B and D demonstrate postoperative segmentation, and how transection localization on the postoperative scan was mapped back to the preoperative CT to determine IRV.

Fig. 1.

CT-based volumetry workflow and estimation of initial residual pancreatic volume. (A) Preoperative volumetry before pancreaticoduodenectomy (PD). The whole pancreatic parenchyma (P) was manually outlined in each axial slice using INFINITT PACS, excluding tumor (T), portal vein (PV), and splenic vein (SV). The surgical transection margin was identified (red arrow) as the reference landmark. (B) Postoperative volumetry after PD. The remnant pancreatic volume was measured by summing parenchymal area × slice thickness (3 mm). The transection margin determined on postoperative CT was retro-matched to the corresponding segment on preoperative CT to estimate the initial residual volume (IRV = 100%). (C) Preoperative volumetry before distal pancreatectomy (DP). (D) Postoperative volumetry after DP using the same manual segmentation protocol.

To ensure methodological consistency and reduce observer-related variability, the entire segmentation process was carried out under the supervision of a senior radiologist with specialized experience in pancreatic imaging, thereby reinforcing the reliability and reproducibility of volumetric measurements.

Statistical analysis

Continuous variables were presented as mean ± standard error (SE), and categorical variables as counts and percentages. Group comparisons were performed using independent t-tests for normally distributed data or Wilcoxon rank-sum tests otherwise. Categorical variables were compared using Chi-square or Fisher’s exact tests as appropriate. Generalized Estimating Equations (GEE) were used to analyze repeated-measure trends in pancreatic volume and C-peptide over time¹⁷. A logistic regression model was applied to analyze the relationship between a binary outcome and a group of predictor variables. The correlation between two groups of continuous variables was analyzed with Spearman’s rank correlation coefficient. All analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA). A two-tailed p value < 0.05 was considered statistically significant.

Results

A total of 135 patients were included, with 90 in the PD group and 45 in the DP group. Malignancy accounted for 66/90 (73.3%) in PD and 20/45 (44.4%) in DP. Clinicopathologic features and perioperative characteristics are summarized in Table 1.

Table 1.

Patient characteristics.

Characteristic	PD (n = 90)	DP (n = 45)	P
Age (mean ± SE), yrs (range)	61 ± 1.25 (25–86)	58 ± 2.06 (20–82)	0.126
Sex (M: F)	56:34	24:21	0.421
BMI (mean ± SE)	23.8 ± 0.37	24.8 ± 0.45	0.154
Preoperative CRT	2	4	0.984
Preoperative jaundice	40 (44.4%)	0	< 0.001***
Preoperative pancreatitis	20 (22.2%)	7 (15.6%)	0.250
Pylorus-preserving PD	72 (80%)	-
Laparoscopic DP Spleen preservation*	- -	22 (48.9%) 8 (17.8%)
Pathology			0.002**
Benign	24 (26.7%)	25 (55.6%)
Malignancy Pancreatic ductal cancer Bile duct cancer Ampullary cancer Duodenal cancer Neuroendocrine cancer Malignant IPMN Metastatic cancer Others	66 (73.3%) 29 7 22 2 - - - 6	20 (44.4%) 11 - - - 4 1 1 3
Preoperative diabetes	24 (26.7%)	21 (46.7%)	0.033*
New onset post-OP DM	15 (22.7%)	8 (33.3%)	0.115
Resolution of pre-OP DM	7 (29.2%)	0 (0%)	0.008**

SE, standard error; M, male; F, female; BMI, body mass index; CRT, chemoradiotherapy; PD, pancreatico-duodenectomy; DP, distal pancreatectomy; IPMN, intraductal papillary mucinous neoplasm; DM, diabetes mellitus; OP, operation

*: Spleen preservation: both splenic vessels and spleen were preserved during operation.

Volumetric changes of the remnant pancreas

In the DP group, the mean preoperative pancreatic volume was 55.51 ± 2.98 ml (range 19.26–121.42 ml), and the initial residual volume (V%) was 67.93% ± 1.97% (range 41.6–91.43%). In the PD group, the mean preoperative pancreatic volume was 58.83 ± 2.63 ml (range 10.39–128.55 ml), with an initial residual volume of 51.20% ± 1.62% (range 13.31–85.76%). No significant difference was noted between groups (p = 0.274).

During follow-up, the remnant pancreas volume increased in DP and decreased in PD. In DP, volume change was 106.25% ± 2.09% at 3 months, 106.62% ± 3.27% at 1 year, and 110.43% ± 4.10% at 2 years (β = 2.97, p = 0.019). In PD, volume change was 80.79% ± 2.86%, 68.67% ± 2.58%, and 65.34% ± 2.98% at 3 months, 1 year, and 2 years, respectively (β = −11.85, p < 0.001). GEE confirmed a significant difference in longitudinal volume trends between groups (β = 5.43, p = 0.014) (Fig. 2C).

Fig. 2.

(A) Hypertrophy in the remnant pancreas after DP, (B) Atrophy in the remnant pancreas after PD, (C) Trend of residual pancreatic volume ratio at 3 months, 1 year, and 2 years postoperatively. Patient numbers at each time point are shown in the table below. a GEE estimate for volume change over time (ref = initial residual volume). b GEE estimate comparing DP vs. PD (ref = PD). (D) Hypertrophic change was common after DP but uncommon after PD. In the DP group, most patients showed hypertrophy at 3 months, 1 year, and 2 years, while PD patients demonstrated a markedly lower rate at corresponding time points (all p ≤ 0.002).

The proportion of patients with hypertrophic changes in DP was 66.7% at 3 months, 71.1% at 1 year, and 73.5% at 2 years, whereas the PD group showed 17.78%, 12.22%, and 10.53%, respectively (Fig. 2D).

Factors associated with volumetric changes in remnant pancreas

Among PD patients, pancreatic duct dilatation was associated with greater pancreatic atrophy. Remnant pancreas volume change in patients with duct dilatation was 77.69%, 58.60%, and 53.89% at 3 months, 1 year, and 2 years, compared to 90.92%, 86.23%, and 85.26% in patients without dilatation (β = −9.82, p = 0.027) (Fig. 3, Supplementary Table 1). Spearman’s correlation analysis confirmed a negative correlation between postoperative duct diameter and volume change at 1 year (ρ = −0.20, p = 0.053) and 2 years (ρ = −0.27, p = 0.02), indicating that larger postoperative duct diameters were associated with more pronounced pancreatic atrophy (Supplementary Fig. 1).

Fig. 3.

Relationship between the residual pancreas volume and P-duct dilatation in the PD group. at postoperative 3 months, 1 year and 2 years. “P-d” means pancreatic duct. “a”, p-value was obtained by two-sample t test; “b”, p-value was obtained by GEE analysis.

In the DP group, the factor, “spleen preservation”, may positively correlate with the degree of hypertrophic changes in the DP group (Supplementary Table 2). We noticed that the volume change of remnant pancreas in the DP patients with and without spleen preservation was 124.55% and 102.74% at 1-year follow-up (p = 0.006), and 120.08% and 107.92% at 2-year follow-up (p = 0.177) respectively (Fig. 4).

Fig. 4.

Relationship between the residual pancreas volume and spleen preservation in the DP group. at postoperative 3 months, 1 year and 2 years. “a”, p-value was obtained by Wilcoxon rank sum test; “b”, p-value was obtained by GEE analysis.

Incidence of new-onset diabetes mellitus

In DP patients, 8/24 (33.3%) non-diabetic patients developed new-onset diabetes mellitus, while none of the 21 preoperative diabetic patients achieved remission. In PD, 15/66 (22.7%) non-diabetic patients developed new-onset diabetes, and 7/24 (29.2%) diabetic patients achieved remission. New-onset diabetes was more frequent in DP, although not statistically significant (p = 0.115). Most cases occurred within the first postoperative year, predominantly between 3 months and 1 year (Supplementary Table 3).

Univariate logistic regression analysis showed no significant association between perioperative variables and new-onset DM in either group (Supplementary Table 4). Neither pancreatic volume nor C-peptide level/C-peptide index was a significant predictor (Supplementary Table 5).

Endocrine function after pancreatectomy

Postoperative endocrine function was compared between PD and DP using C-peptide index (CPI) (Fig. 5A). At the 3-month follow-up, the DP group exhibited a significantly lower CPI compared with PD (p = 0.004), whereas no statistical difference was observed preoperatively or at 1- and 2-year assessments (p = 0.419, 0.231, and 0.207, respectively). To account for inter-individual variation, C-peptide recovery was further evaluated using CPI ratio (post-operative CPI relative to pre-operative baseline = 100%) (Fig. 5B). Both groups showed an initial decline at 3 months followed by partial recovery. GEE analysis demonstrated no significant difference in overall recovery trend between PD and DP (surgery type p = 0.12; time × surgery interaction p = 0.56). Although CPI reduction appeared greater in the DP group at individual time points, none reached statistical significance.

Fig. 5.

Postoperative endocrine function and morphologic-functional comparison between PD and DP. (A) C-peptide index (CPI) was significantly lower in DP at 3 months (p = 0.004) but showed no difference at other time points. (B) CPI ratio demonstrated initial decline at 3 months with partial recovery in both groups, with no significant difference in overall recovery trend (GEE: surgery type p = 0.12; time × surgery interaction p = 0.56). (C, D) PD showed greater volume loss with greater CPI recovery, while DP demonstrated less pronounced volume reduction with stabilization after 3 months and similar endocrine recovery dynamics.

Morphologic–functional comparison demonstrated different patterns between structural adaptation and endocrine preservation (Fig. 5C and D). In PD, pancreatic volume decreased substantially after surgery and remained low through 2 years, whereas CPI ratio gradually improved despite minimal volumetric recovery. In DP, volume reduction was less pronounced and stabilized after 3 months, with endocrine function showing similar recovery dynamics.

To account for differences in residual pancreatic volume, endocrine recovery was further assessed using C-peptide index normalized to remnant volume (CPI/volume ratio). GEE analysis showed no significant change in CPI/volume ratio over time (β = 2.3 ± 3.7, p = 0.531), indicating relatively stable endocrine efficiency per unit pancreatic tissue postoperatively. However, the PD group demonstrated better functional recovery than DP (treatment β = 126.9 ± 36.9, p < 0.001), and the time × surgery interaction was significant (β = 28.4 ± 14.5, p = 0.049), suggesting a favorable improvement trajectory in PD even after adjusting for remnant volume (Table 2).

Table 2.

Relationship between surgical treatment (PD or DP) and the ratio of the change of C-peptide index to the volume of remnant pancreas *.

Variable	GEE
	β ± SE	p-value
Treatment
DP	Ref.
PD	126.9 ± 36.9	< 0.001***
Time	2.3 ± 3.7	0.531
Treatment*Time (interaction)
DP	Ref.
PD	28.4 ± 14.5	0.049*

^* The ratio of the change of C-peptide index to the volume of remnant pancreas was obtained by C-peptide index ratio (%) ÷ residual pancreas volume (%).

Discussion

This study provides a long-term, quantitative evaluation of postoperative pancreatic remodeling and endocrine dynamics following pancreaticoduodenectomy (PD) and distal pancreatectomy (DP). By simultaneously assessing volumetric change and endocrine function over a two-year period, our findings reveal that structural regeneration and functional recovery of the pancreas do not necessarily progress in parallel. Specifically, remnant pancreatic atrophy was observed predominantly after PD, whereas hypertrophic compensation occurred after DP. Despite these contrasting morphological trajectories, both groups showed partial restoration of insulin secretion over time, indicating that β-cell functional recovery may occur through mechanisms distinct from parenchymal volume expansion.

Similar to other previous reports, the remnant pancreas in the PD group in this study also atrophied, with pancreatic duct dilatation being a major factor associated with volume atrophy^18–21. Pancreaticojejunostomy anastomosis stricture, which results in the dilatation of the pancreatic duct, can lead to postoperative acute and chronic pancreatitis in humans and is associated with subsequent morphological and functional changes^22,23. On the other hand, we found that most of the remnant pancreas (73.5%) in the DP group exhibited hypertrophic changes at a 2-year follow-up, especially in patients who underwent spleen-preserving distal pancreatectomy. In our center, spleen preservation is performed using the Kimura technique²⁴, maintaining both the splenic artery and vein. Since distal body/tail benign or low-risk malignant lesions are typically selected for spleen-preserving resection, a portion of the proximal pancreatic body is often retained and remains perfused by splenic vessels, which may contribute to better remnant perfusion and support parenchymal maintenance. Besides the blood supply issue, experimental studies have suggested that spleen-derived mesenchymal cells may enhance pancreatic regeneration in animal models^25–27, but this mechanism has not been validated in humans and should be interpreted cautiously. Therefore, larger cohorts are required to clarify whether spleen-preserving DP confers any functional advantage to the remnant pancreas.

While liver regeneration shows a positive correlation between volume recovery and functional restoration following hepatectomy²⁸, our study revealed that volume change after pancreatectomy did not correlate with the endocrine function change. C-peptide is commonly used to evaluate endogenous insulin secretion by the β cells in the pancreas, indirectly reflecting β-cell function, and also less affected in insulin-treated patients^29,30. Because circulating C-peptide levels are influenced by concurrent glucose concentration, β-cell function was further evaluated using the C-peptide index (CPI)¹⁶. We observed that while the PD group’s remnant pancreas underwent progressive atrophy, their C-peptide levels and CPI paradoxically improved over time. Similarly, Jung et al. reported no significant difference in C-peptide level according to severity of atrophy³¹. This functional improvement despite volume atrophy can be explained by the remarkable plasticity of β-cells, which can significantly vary their insulin secretory capacity based on metabolic demand³². Studies have shown that individual β-cells can increase their insulin output by up to 100% under conditions of increased needs through enhanced secretion rather than cellular proliferation or hypertrophy³³. These observations demonstrate that preservation of endocrine function depends on both the volume and anatomical location of preserved tissue, while functional recovery relies more on the adaptive capacity of remaining β-cells than on volume changes.

Despite preserving significantly more pancreatic volume than PD (p = 0.027), DP patients experienced greater C-peptide index decline (p = 0.004). Our data also revealed that PD group demonstrated better endocrine functional recovery than DP. This finding can be explained by the anatomical distribution of islet cells - the tail and body regions contain approximately 2-fold higher islet density compared to the head region, making the loss of endocrine tissue more significant in DP despite similar total volume resection^34,35. Animal models further support this regional functional asymmetry, with dorsal-bud–derived tail islets exhibiting higher insulin secretory capacity than ventral-bud-derived head islets³⁶. Collectively, these findings may support the idea that distal resection leads to a greater metabolic impact than proximal resection. Therefore, for benign or low-risk malignant lesions located in the pancreatic neck or proximal body, central pancreatectomy with preservation of the pancreatic tail should be considered to minimize postoperative endocrine dysfunction.

The development of new-onset diabetes mellitus (NODM) and the potential for diabetes remission after pancreatectomy represent important postoperative outcomes in our study. Previous studies have reported NODM rates of 14.5–22.2% after PD³⁷, with diabetes remission occurring in 30.7–65% of patients with preexisting diabetes^38–40. In our cohort, we observed similar patterns in the PD group, with a NODM rate of 22.7% and diabetes remission in 29.2% of patients with preexisting diabetes. However, consistent with previous studies^41,42, our DP group showed notably different outcomes, with a higher NODM rate (33.3%) and no cases of diabetes remission, though the difference was not statistically significant due to the small sample size. The higher incidence of NODM and lack of diabetes remission in our DP group, despite greater volume preservation, can be explained by two factors: First, the β-cell density and distribution are approximately twice as high in the pancreas tail compared to the head region^34,35. Second, after PD, which involves bypassing the duodenum and proximal jejunum, patients undergo physiological and anatomical akin to those seen after Roux-en-Y gastric bypass (RYGB) surgery. RYGB is known to enhance metabolic profiles and reduce body weight by increasing incretin hormones, such as glucagon-like peptide-1 (GLP-1). This rise in incretin hormones may also improve glycemic control in PD patients^43,44. As a result, DP has a more substantial impact on insulin secretion, leading to a higher incidence of endocrine dysfunction.

This study has several limitations. First, it was conducted at a single institution with a relatively small sample size, which may introduce bias. In addition, our inclusion criteria required patients to have adequate longitudinal follow-up with CT volumetry and metabolic assessments for at least one year—and excluded those with tumor recurrence within two years—patients with advanced or aggressive malignant disease were less likely to be included. This selection process may have introduced selection bias, resulting in a cohort that was healthier and not fully representative of the entire PD or DP population. Second, data on insulin resistance (e.g., HOMA-IR) were not available, limiting further exploration of mechanisms contributing to postoperative diabetes. Third, dietary intake, digestive/absorptive capacity, and nutritional parameters such as postoperative body weight were not systematically collected. Given that gastrointestinal reconstruction differs substantially between PD and DP, variations in nutritional status may have affected pancreatic volume adaptation and insulin secretion, and the absence of these data prevents full adjustment for these potential confounders.

In conclusion, Pancreatic volume and endocrine function recover independently after pancreatectomy. Despite greater volume preservation, DP patients experience more endocrine dysfunction than PD due to higher islet density in the distal pancreas. Pancreatic duct dilatation predicts atrophy in PD. Surgical strategy should prioritize islet preservation rather than total volume. For benign pancreatic lesions, central pancreatectomy with tail preservation is recommended to optimize metabolic outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Wei-Hsun Lu, MD, MS: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft preparation, Writing – review and editingHong-Ming Tsai, MD: Methodology, Software, Validation, Visualization, Writing – review and editing (specifically provided expertise in CT volumetry calculation and medical image processing)Ting-Kai Liao, MD: Data curation, Investigation, Resources, Writing – review and editing (assisted in patient recruitment and data collection)Ping-Jui Su, MD: Data curation, Investigation, Resources, Writing – review and editing (assisted in patient recruitment and data collection)Chih-Jung Wang, MD: Data curation, Investigation, Resources, Writing – review and editing (assisted in patient recruitment and data collection)Ying-Jui Chao, MD: Data curation, Investigation, Resources, Writing – review and editing (assisted in patient recruitment and data collection)Yan-Shen Shan, MD, PhD: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Validation, Writing – review and editing (initiated the study concept, supervised the research design and analysis, and oversaw manuscript preparation and submission).

Funding

The grant was supported by Taiwan Ministry of Science and Technology (MOST 108-2321-B-006-014), Ministry of Healthy and Welfare (MOHW 107-TDU-B-212–114026 A), and HSU-YUAN Education Foundation (HY-2018-001).

Data availability

The datasets used in the study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

IRB approved by the Institutional Review Board of the National Cheng Kung University Hospital with approval number of B-ER-108-312.

Consent to participate

Given the retrospective nature of this study and the use of de-identified patient data, the Institutional Review Board of the National Cheng Kung University Hospital waived the requirement for informed consent for study participation. All patient data were completely anonymized prior to analysis, and no identifying information was included in the study dataset or manuscript. The waiver of informed consent was approved as part of the ethical review process under approval number B-ER-108-312.