Abstract
Background/Aim: Pancreatic intraepithelial neoplasia (PanIN) is the most common precursor lesion to pancreatic adenocarcinoma (PDAC). Yet no criteria to quantify patients at risk for progression to PDAC with PanIN exist. Platelet to lymphocyte ratio is an inflammatory marker that has been associated with overall survival in patients with invasive malignancies including pancreatic cancer. Preoperative sarcopenia has been linked to more aggressive diseases in pancreatic neoplasms. We aimed to assess a relation between PLR and sarcopenia as predictors for tumor progression in patients undergoing pancreatic resection for IPMN. Patients and Methods: We retrospectively reviewed 102 patients (46 females, 56 males) who underwent pancreatic resection for PanIn. PLR was calculated and quantified using a cutoff of 110, sarcopenia was quantified using the skeletal muscle index (SMI) on preoperative abdominal imaging. Both were co-evaluated with additional demographic, clinical, pathological, and imaging data for possible correlation with PanIN associated PDAC. Results: PLR was significantly elevated in patients with PanIN - associated PDAC (p=0.006). In the multivariate analysis, invasive carcinomas were significantly more prevalent in patients with PLR above 110 (OR=4.06, 95%CI=3.91-4.12, p=0.04). Patients with elevated PLR had a two-times higher risk to die in the postoperative period (HR=2.26, 95%CI=1.04-2.21, p=0.001). Patients with elevated PLR, preoperative jaundice and sarcopenia were the most likely to have PanIN-associated PDAC (OR=3.48, 95%CI=2.98-8.41, p=0.02). Conclusion: PLR is an independent predictive marker for the presence of PanIN associated invasive carcinoma.
- Pancreatic intraepithelial neoplasia
- pancreatic carcinoma
- platelet
- lymphocyte
- platelet/lymphocyte ratio
- sarcopenia
- jaundice
Pancreatic intraepithelial neoplasia (PanIN) is the most common precursor lesion to pancreatic adenocarcinoma (PDAC) (1, 2). PanINs are microscopic mucinous papillary lesions (<5 mm) which progress to cancerous lesions using the adenoma–carcinoma sequence as well as epithelial-to-mesenchymal transition (EMT) (3). PanIN lesions are thought to develop by stepwise accumulation of genetic and epigenetic alternations in a progression from low-grade PanIN (LG-PanIN) to high-grade PanIN (HG-PanIN) and finally to pancreatic ductal adenocarcinoma (PDAC) (4, 5). EMT regulators that promote cancerous development in PanIN are reportedly the same that have been associated with tumor cachexia and muscle loss in other malignant disease (5, 6).
PanIN cannot always be identified by abdominal imaging or other noninvasive diagnostic techniques. In individuals with a biopsy proven PanIN surgical removal is indicated (7, 8). There are still PanINs that do not progress to invasive cancer; hence these patients are overtreated with a pancreatic resection. Until now, no parameter to distinguish between patients who would require resection of proven PanIN and patients who might not benefit from resection exists.
The development of pancreatic cancer may lead to impairment of the patient’s immune system through systemic inflammation (9). Increased neutrophil-to-lymphocyte ratio (NLR) and increased platelet-to-lymphocyte ratio (PLR) are inflammatory markers that reflect this immune response and have been previously correlated with poor overall survival in patients with other malignancies (10, 11). Recently, both markers were identified as independent predictors for the presence of malignancy in pancreatic cystic lesions (12).
In the recent past, low skeletal muscle mass has been recognized as a surrogate parameter for frailty (13, 14). Low skeletal muscle mass reportedly occurs independently from the patients’ weight or body mass index and is distinct from cancer-related weight loss or cachexia. Studies found a relation of low skeletal muscle mass and worse postoperative outcomes (15, 16).
The primary aim of the study was to verify the hypothesis that increased NLR and PLR in preoperative blood samples from diagnosed PanIN patients are associated with the presence of PDAC. As secondary aim PLR and NLR were evaluated as predictors for patient survival along with clinicopathological factors that have already been associated with patient survival. To answer this, a retrospective study on a series of resected PanIN patients was performed in our institution to determine the predictive value of NLR and PLR in association with additional clinical and radiological parameters.
Patients and Methods
In the present study patients with PanIN were selected. To verify PanIN a fine needle aspiration biopsy of a suspected mass in the pancreas was performed. The biopsies were carried out by radiological specialists, fixed in formalin, and sent to the pathologist for routine histopathological work up. Histological analysis was carried out as a standard diagnosis. The patients were scheduled for resection as soon as the results were available, and the patient agreed to the presented treatment concept. The study cohort comprises a selection of these patients as described below.
Patients underwent pancreatic resection after written informed consent. Most patients underwent Whipple procedure [Whipple procedure (96%) or distal pancreatectomy (4%)]. Specimens were resected and sent to frozen section analysis and processed histopathologically after frozen section (resection margin) analysis. No histopathological analysis beyond the routine histopathological workup was carried out for the presented study. Patients who were classified as having a malignancy in the definitive specimen were classified as having progressed to malignancy in the presented analysis.
Records of 198 patients who had reported PanIn lesions in their pathology reports and underwent pancreatic resection between 2006-2016 for pancreatic lesions were scanned. Patients with a history of prior, synchronous, or metachronous malignancy; prior transplantation; autoimmune disease; human immunodeficiency virus infection; and treatment using immunosuppressive agents were excluded from the analysis. After exclusion of these patients, the cohort consisted of 102 patients.
Patient-specific data (demographics, pathological, and clinical) were collected and analyzed. Weight loss was defined as unintentional loss of more than 5 kg reported at presentation for operation and jaundice as elevated total bilirubin blood level of more than 3 mg/dl.
NLR was calculated as the total count of neutrophils divided by the total count of lymphocytes and PLR as the ratio of the total count of platelets divided by the total count of lymphocytes (all values in microliter). In a healthy individual, NLR less than 2 and PLR less than 80 can be considered normal. For our analysis we used previously applied cutoffs for occurrence of PanIN associated invasive carcinoma in pancreatic patients of NLR of 4 and PLR of 110 (17). Using the Osirix DICOM viewer (Pixmeo, SARL, Bernex, Switzerland), low skeletal muscle mass was initially assessed by measuring the skeletal muscle index (SMI) at the level of L3 in accordance with previous publications (18, 19). As previously described, measurements were performed in a semi-automated fashion with manual outlining of the skeletal muscle borders with the density threshold setting between –30 and 110 Hounsfield Units (HU) to exclude vascular and fatty infiltration areas from the volumetric calculations. SMI was normalized for height (body surface area (m2): (height(cm) x weight(kg))/3,600).
Statistical analysis. Continuous and categorical variables were reported as medians with interquartile ranges (IQR) and as whole numbers and percentages, respectively. The distributions of categorical and numerical variables between independent groups were compared using Fisher exact test and a Mann-Whitney U-test. Statistically significant variables in the univariate analysis were subjected to multivariable logistic regression model with backward elimination method (likelihood-ratio test). Odds ratio (OR) and 95% confidence interval (CI) were calculated to determine the probability of PanIN associated adenocarcinoma. Overall survival was analyzed using the Kaplan-Meier method, and differences in survival were evaluated with the log-rank test. The association of relevant clinicopathological variables with postoperative mortality was computed using Cox proportional hazards models; backward stepwise selection was used to identify variables for the multivariable Cox proportional hazards model. Results were reported as hazard ratios (HR), where appropriate, with 95%CI. The impact of low skeletal muscle mass was evaluated both as a continuous and a categorical variable. As previously reported and validated, to obtain the specific sex categorical cutoff value for low skeletal muscle mass, optimum stratification was assessed through a series of sensitivity analyses and low skeletal muscle mass was defined in categorical analyses as the lowest quartile. The SMI cutoffs to define low skeletal muscle mass were 22.1 cm2/m2 in females and 33.7 cm2/m2 in males. SMI entered univariate and multivariate analysis as described above. A p-value of less than 0.05 was considered to be statistically significant. For statistical analysis, SPSS 26.0 (IBM, Chicago) was used, images were generated using SciStat (www.scistat.com, MedCalc, Ostend, Belgium).
Results
The baseline clinicopathological characteristics of 102 patients who met the study inclusion criteria are summarized in Table I. Median patient age was 63 years (range=54-75) and 65% (n=66) were males. Median BMI was 25.9 (23.4; 28.8). Over sixty percent of patients (n=68) had a Charlson Comorbidity Index >6 reflecting severe comorbidities. Most patients (53%) were classified as high-risk according to the preoperative ASA score (ASA 3-4; n=54). Median length of stay was 23 (16; 31) days, 8.4% of patients experienced postoperative complications above Clavien Dindo Grade III (n=9) indicating a safe procedure. The median preoperative hemoglobin (Hb) levels in the whole patient set were 13.9 g/dl (11.8; 14.2). Most patients showed high-grade dysplasia in their specimen (43%) and most specimens were obtained by the Whipple procedure (96%). Median operative time was 4.7 hours (3.2-8.7) and 32% (n=33) received intraoperative blood transfusions. Median NLR and PLR were 2.7 and 167.2 respectively indicating an elevation in both parameters in the whole patient set. Sarcopenia according to the SMI was present in 25% of patients (n=26), median SMI was 60.5 (50.95-69.57) and patients had sarcopenic obesity (8.6%) (Table II).
Baseline clinicopathologic characteristics of all included patients.
Patients were compared according to the histological grading in their resected pancreas. Twenty-three patients already showed progression to malignancy after operation. Preoperative weight loss was much more prevalent in patients whose disease already had progressed to PDAC. Results in the comparison of NLR; PLR and CA 19-9 suggested possible associations with progression to malignancy.
Most patients had a HG-PanIN (n=44), only 23 of resected PanIn had already progressed to PDAC. Preoperative weight loss was much more prevalent in patients whose disease had already progressed to PDAC, of these patients the majority (55.6%) reported preoperative weight loss compared to 48.3% and 41.3% in LG and HG PanIN (p=0.05). PDAC patients also showed higher rates of duct dilation (33.3% vs. 20.7 and 25%; p=0.03) and had significantly higher serum levels of CA 19-9 as compared to LG and HG PanIN (p=0.001). The preoperative NLR as well as the PLR were significantly higher in patients who had progressed to PDAC as compared to patients who still had LG or HG PanIn (p=0.03).
Patients who presented with preoperative jaundice had a 30% higher risk to already have progressed to PDAC compared to patients without jaundice (OR=1.27, 95%CI=0.45-3.6; p=0.01). Higher NLR >4 was associated with a near 3-times higher risk of progressed disease (OR=2.67, 95%CI=1.63-11.24; p=0.03). Patients with PLR above 104 showed an even 4 times higher risk to have PanIN associated PDAC (OR=3.71, 95%CI=1.04-7.09; p=0.03). Other preoperative factors like preoperative sarcopenia, saropenic obesity or solid component imaging did not show any significant relations to progression to PDAC (Table III).
Univariate analysis confirmed a possible relation between neutrophil to lymphocyte ratio, platelet to lymphocyte ratio and jaundice with the progression to malignancy in patients with PanIN.
In multivariate analysis the progression to PDAC was used as an outcome parameter. Factors tested were present preoperative jaundice, main pancreatic duct dilation >5 mm, patients age, preoperative weight loss, intraoperative transfusions, preoperative sarcopenia, CA 19-9 elevation, NLR >4 as well as PLR >110. As expected, preoperative grading divided into low/intermediate grading, and high-grade dysplasia was significantly associated with the development of PanIN associated PDAC. Out of independent parameters, preoperative jaundice, a PLR >110 and the presence of preoperative sarcopenia remained significant predictors for the presence of a malignancy (p=0.041, p=0.041 and p=0.02, respectively) (Table III). The Area under the receiving characteristic (AUC) for all parameters was 0.709 with a specificity of 60% and a sensitivity of 67% (95%CI=0.583-0.836; p=0.001). The addition of dysplasia grading to sarcopenia, jaundice and PLR >110 did ameliorate the AROC to 0.839 (95%CI=0.771-0.894; p=0.024) (Figure 1).
Area under the receiving characteristic (AUC) for sarcopenia, jaundice and PLR >110 was 0.709 with a specificity of 60% and a sensitivity of 67% (95%CI=0.583-0.836, p=0.001, addition of dysplasia grading to sarcopenia, jaundice and PLR >110 did ameliorate the area under the curve (AUC) to 0.839 (95%CI=0.771-0.894, p=0.024). Sarc_Jaund_PLR: Sarcopenia, jaundice and PLR >110 present; SJPLR_grading: sarcopenia, jaundice and PLR >110 present.
Median postoperative survival of patients without jaundice, sarcopenia and normal PLR was significantly higher compared to patients with all 3 co-variants (8.3 years 95%CI=7.4-9.3 vs. 3.4 years 95%CI=1.6-5.2; p=0.001). Addition of several co variants lessened survival rates gradually from 8.4 years estimated median postoperative survival in patients with jaundice to 5.4 years estimated median postoperative survival in patients with jaundice and PLR >110 to 3.4 years in patients with jaundice, PLR >110 and sarcopenia (Figure 2). On Cox proportional hazard analysis using survival as endpoint presence of sarcopenia and PLR >110 remained independently associated with lower survival (p=0.008 and p=0.045, respectively) compared to other co variants (age, BMI, dysplasia, duct dilation, preoperative jaundice, NLR, T stage, CA 19-9) (Table IV).
Median postoperative survival deteriorated with addition of every co variant from 9.1 years estimated median postoperative survival in patients without any co factor to 8.4 in patients with jaundice to 5.4 years estimated median postoperative survival in patients with jaundice and PLR >110 and to 3.4 years in patients with jaundice, PLR >110 and sarcopenia. Blue line: patients without any co factor, orange line: patients with jaundice, green line: patients with jaundice and PLR >110, green tickled line: patients with jaundice, PLR >110 and sarcopenia.
Multivariate analysis that used the progression to PDAC as outcome parameter showed that preoperative sarcopenia, jaundice as well as PLR have a statistically significant association.
Discussion
This is the first study combining different parameters in order to depict possible existing malignancy in patients undergoing surgery with PanIN lesions.
Patients who undergo pancreatic resection are a widely diagnosed patient group with many available preoperative information. We sought to use this information to combine different parameters in order to predict malignant progression. Sarcopenia has been described as a parameter for patients immediate and long-term postoperative survival after pancreatic resection before but has not been evaluated in the setting of malignant progression in precursor lesions like PanIN (20). The latter has been associated with poor prognosis in patients with PDAC before. It has been associated with lower survival as with higher recurrence rates. Meta-analysis on a large patient cohort with PDAC by Li et al. showed a high relation between PLR and overall patient survival (21). In the presented study patients with PLR >110 even had a worse survival in the presence of jaundice and/or sarcopenia. A recent meta-analysis by Zhou et al. also confirmed that low PLR was associated with longer progression free survival when compared with high PLR in patients with pancreatic cancer. This underlines the influence of PLR on the development of malignancy in patients with precursor lesions in the presented cohort (22).
Ferri and colleagues reported an association between preoperative jaundice, elevated CA199 levels and the identification of cancerous lesions in the pancreas. This was also true in our patients. Jaundice was significantly associated with malignancy in the presented patients (Table III) (10, 23, 24).
Until now most studies only target patients with previously known malignancies (25). In this study patients with precursor lesions were included and progression to malignancy was used as an outcome parameter. Zhou et al., in a review, suggested that this progress is mediated by EMT and can be predicted using EMT determinants (26). Krebs et al. found that EMT transcription factors are expressed in metastatic PDAC and play a role in pancreatic tumorigenesis (27). Methods to determine EMT and its transcription factors are expensive and only available with native pancreatic tissue. All parameters used for the presented analysis are assessed routinely preoperatively and by that do not cause further costs.
The mechanism by which high PLR affects poor survival of cancer patients remains unclear, but may be partly explained through an inflammatory process caused by cancer cells (28). The presence of tumor cells affects platelets and may lead to cancer-induced thrombosis (29). Platelets also release a number of tumor growth factors, which support angiogenesis and metastasis (17). Lymphocytes play a large role in cancer immune-surveillance, which prevents tumor development. Cancer-related inflammation causes suppression of antitumor immunity by recruitment of regulatory T cells and activation of chemokines, which results in tumor progression (30). Recently Takeuchi et al. reported a relation between lymphocyte to monocyte ratio and lower survival after pancreatic resection in pancreatic cancer (31). According to the results of this study, preoperative high PLR was a significant risk factor for the development of PDAC even independently of tumor grading.
Sarcopenia has also already been associated with poorer prognosis in patients with PDAC. It has often been combined with different preoperative scores (13, 15, 19); however, sarcopenia has not been used to assess progression to PDAC thus far.
This study has several limitations. First, it was conducted at a single institution with a small sample size. Second, we included patients who underwent Whipple procedure as well as patients who underwent distal pancreatectomy. Both procedures are not entirely comparable in risk and postoperative course. Third, factors that affect PLR value, like liver function were not taken into account. Fourth, the study was performed under a retrospective design and therefore results have to be verified in a prospective approach.
In conclusion, we found that a combination of PLR, sarcopenia according to the SMI and preoperative jaundice is useful to determine a higher risk for malignancies in patients with diagnosed PanIN. In the presence of these factors, patients should be resected quickly and are likely to have been progressed to PDAC.
Footnotes
Authors’ Contributions
Doris Wagner, Valerie Wienerroither, Johannes Haybaeck: writing and drafting manuscript. Tarik Bajric, Andrea Tomberger: data acquisition, measurements. Peter Schemmer, Hans Jörg Mischinger: drafting and editing the manuscript. Peter Kornprat: conceptualization of study, drafting the manuscript.
Conflicts of Interest
The Authors do not have any conflicts of interest to declare.
- Received October 30, 2021.
- Revision received January 12, 2022.
- Accepted January 20, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.







