Diagnostic accuracy of cfDNA levels in gallbladder cancer: A study using qPCR & threshold evaluation

Materials & Methods

This case-control study was conducted with participants recruited consecutively from the departments of Gastroenterology, Surgical Oncology, Gastrointestinal Surgery, Radiation Oncology, and Pathology, Sawai Man Singh Medical College (SMS), Jaipur, Rajasthan, India after obtaining the ethical approval from the Institutional Review Board. A written informed consent was obtained from all participants. All demographic and clinical details, laboratory investigations, and study participants imaging were meticulously recorded.

The study included 42 newly diagnosed adult individuals affected with GBC. Diagnoses were established using a combination of clinical presentation, imaging such as ultrasound (USG), computed tomography (CT), magnetic resonance imaging (MRI), and Fine Needle Aspiration Cytology (FNAC) or biopsy wherever feasible. All cases were identified as suspected for malignancy when USG, with CT and MRI findings were consistent for GBC in 31 cases. FNAC confirmed malignancy in 23 cases, and a biopsy was conducted in four cases to establish the diagnosis. Due to the advanced stage and critical condition of many patients, histopathological confirmation was not possible in all cases. Individuals with prior chemotherapy or radiotherapy, those with significant clinical disorders, or those with inadequate samples were excluded from the study. Additionally, 15 healthy individuals, matched for age and sex to the cases, were included as controls.

qPCR and standard curve

The cfDNA was quantified using qPCR on the CFX96 Real-Time System, following the Applied Biosystems protocol. The qPCR reactions were performed using SYBR Green Supermix with specific primers (forward: GTGCACCTGACTCCTGAGA; reverse: CCTTGATACCAACCTGCCCAG) ^13,14 under the following cycling conditions: an initial denaturation at 95°C for 9 min, followed by 35 cycles of 95°C for 30 sec, 55°C for 30 sec, and 72°C for 30 sec. A standard curve (Fig. 1; A and B) was established by plotting Ct values against the log₁₀ of known DNA concentrations, modelled using polynomial regression (degree 3), and R² was calculated ¹⁵ .

Close

Fig. 1.

(A) Regression analysis showing a comparison of Ct values and log10 DNA concentration between cases and controls. (B) Receiver operating characteristics (ROC) curve for diagnostic performance evaluation of cfDNA.

Export to PPT

The diagnostic capabilities of cfDNA were further evaluated by plotting receiver operating characteristic (ROC) curves (Fig. 1B) and calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) at various cfDNA concentration thresholds. Four key thresholds (75.5 ng/ml, 130 ng/ml, 188 ng/ml, and 372.92 ng/ml) were identified, each offering different clinical advantages based on the balance between sensitivity and specificity. Precision-recall curves were also generated to visualise the trade-offs between identifying true cases and avoiding false positives.

Results

The demographic and clinico-pathological characteristics of the affected individuals and control groups are summarised in table I. The age range of cancer affected individuals varied from 31 to 75 yr, with a mean age of 55.07±12.14 yr, whereas the control group had a mean age of 19.38±0.68 yr. The gender distribution in the cancer group showed a male-to-female ratio of 16:25, which was matched in the control group.

Most of the cancer affected individuals were at stage IV (95.2%), with a few at stages II and III. Lymph node involvement and wall thickness >5mm were present, respectively in 52.4 per cent and 42.86 per cent of cancer affected individuals. Additionally, jaundice was observed in 69 per cent, pruritus in 35.71 per cent, and ascites in 38.1 per cent of the affected individuals. Our study’s reliance on imaging and FNAC for GBC diagnosis highlights the real-world challenges of managing advanced-stage patients, where histopathological confirmation may not be feasible. While our study focused on critically ill individuals with confirmed or highly suspected malignancy, future research should aim to include benign conditions to provide a comprehensive assessment of cfDNA’s diagnostic utility.

Analysis of tumour markers revealed notable trends. Alkaline phosphatase (ALP) was elevated in 64.29 per cent of affected individuals, making it the most common marker, potentially indicating liver metastasis. In contrast, Alpha-fetoprotein (AFP) was elevated in only 7.14 per cent of affected individuals, being the least common marker. Carcinoembryonic antigen (CEA) levels were elevated in 14.29 per cent of affected individuals.

Demographic and clinico-pathological association

The study cohort included GBC affected individuals with a mean age ± standard deviation of 55.07±12.14 yr, compared to 19.38±0.68 yr in the control group. The age difference was not statistically significant (P=0.409). Gender distribution in the cancer group consisted of 17 males and 25 females, while the control group included seven males and eight females, showing a significant association between cfDNA levels and gender (P=0.045). The statistical analyses of cfDNA levels are presented in table I.

Tumour staging revealed that 95.2 per cent of GBCs were at stage IV, with no significant association between staging and cfDNA levels (P=0.132). Tumour wall thickness greater than 5 mm was observed in 52.4 per cent of cases. Although cfDNA levels were higher in this group (mean± standard deviation: 1420.8 ng/ml ± 866.1) compared to cases with thickness ≤5 mm (802.9 ng/ml ± 1405.3); the association was not statistically significant (P=0.21).

Additionally, jaundice was present in 69 per cent of cases, and a significant correlation between jaundice and cfDNA levels was observed (P=0.035). A stratified analysis of cfDNA levels by jaundice status revealed significantly higher cfDNA levels in jaundiced individuals (1911.5±1532.2 ng/ml) compared to non-jaundiced individuals (368.1±297.4 ng/ml), with a statistically significant difference (P=0.011). This finding highlights the potential confounding effect of elevated bilirubin on cfDNA levels.

The association between cfDNA levels and both metastasis and tumour size (measured by wall thickness) were evaluated using the Mann-Whitney U test. Results showed that cfDNA levels were significantly elevated in affected individuals with metastasis (P=0.044), suggesting a correlation between cfDNA levels and disease spread.

Although the analysis between age groups (>50 vs. ≤50) and lymph node involvement did not reach statistical significance, trends indicated higher cfDNA levels in younger patients and those with lymph node involvement, potentially reflecting a more aggressive disease course in these subgroups. Box plots (Fig. 2) illustrate cfDNA levels by gender, jaundice presence, and tumour size, supporting cfDNA’s role as a potential biomarker of GBC severity.

Close

Fig. 2.

The box plots illustrate the relationship between cfDNA levels and various demographic and clinico-pathological characteristics. (A) by age group; (B) gender; (C) jaundice and (D) wall thickness.

Export to PPT

Distribution analysis of cfDNA levels

The relationship between Ct values and cfDNA concentration was modelled (Fig. 1A) using polynomial regression (degree 3), yielding R² values of 0.6621 for cases and 0.9999 for controls. These high R² values demonstrate our qPCR assay’s reliability. The QQ plots (Fig. 3) showed deviations from normality in both groups, with a more pronounced deviation in cancer patients, further supporting cfDNA as a biomarker.

Close

Fig. 3.

The QQ plots compare the cfDNA distributions for the cases (A) and controls (B) against a normal distribution.

Export to PPT

Diagnostic accuracy of cfDNA

cfDNA levels were markedly elevated in the cancer group (721.18 ng/ml) compared to the control group (81.15 ng/ml, P<0.005). These findings underscore the potential of cfDNA as a reliable biomarker for distinguishing individuals affected with GBC from controls.

The ROC curve (Fig. 1B) for cfDNA levels in GBC affected individuals was plotted, and the AUC was calculated. An AUC of 0.94 was obtained, indicating that our model had a 94 per cent chance of accurately distinguishing between GBC affected individuals and controls.

Precision-recall curve analysis (Fig. 4; A, B and C) was used to evaluate the diagnostic capability of cfDNA levels in distinguishing between healthy controls and cancer affected individuals. We identified four key cfDNA thresholds (75.5 ng/ml, 130 ng/ml, 188 ng/ml, and 372.92 ng/ml) by expanding the PR curves (Fig. 4A), each offering distinct clinical benefits. The 75.5 ng/ml threshold, with high sensitivity (97.6%), is ideal for initial screenings, while the 130 ng/ml threshold provided balanced sensitivity (75%) and specificity (70%), suitable for follow-up testing (Table II). The 188 ng/ml threshold excelled in precision (92%) and specificity (100%), making it optimal for diagnostic confirmation. The 372.92 ng/ml threshold maximizes specificity (90%), which is ideal for confirmatory testing (Fig. 4B).

Close

Fig. 4.

(A)The expanded precision-recall curve shows the overall performance of the model across multiple thresholds, illustrating the trade-offs between precision and recall. (B) The precision-recall curves at specified thresholds (75.5 ng/ml, 130 ng/ml, and 188 ng/ml) displays their effect on the model’s ability to distinguish between cancer affected individuals and healthy controls. (C) This figure demonstrates the precision of the model at various cfDNA concentration thresholds.

Export to PPT

Table II. Table shows selected thresholds and its clinical significance

Threshold (ng/ml)	Sensitivity (%)	Specificity (%)	Significance
75.5	97.6	40	Initial screening; critical to detect as many cases as possible
130	75	70	Follow up testing; balanced approach to detect cases and minimize false positives
188	59.5	100	Diagnosis confirmation; minimizing false positives is crucial
372.92	50	90	Final confirmatory testing; ensuring high accuracy in positive results

Precision at different thresholds: 75.5 ng/ml, 130 ng/ml, 188 ng/ml, and 372.92 ng/ml were calculated as 82 per cent, 79 per cent, 86 per cent, and 81 per cent, respectively (Fig. 4C). The bar graphs of PPV, NPV, sensitivity, and specificity at each threshold, visually illustrate the diagnostic accuracy of cfDNA levels and provide a clear representation of the test’s performance (Supplementary Figure).

Supplementary Figure

Download

Discussion

GBC is highly prevalent in northern and northeastern States, where it ranks among the top cancers ^2,3 . The disease is often detected at an advanced stage, leading to poor life expectancy, lower progression-free survival (PFS), and overall survival (OS) rates ¹⁵ . Its aggressive nature and resistance to conventional chemotherapy make early detection and intervention crucial. Association with chronic inflammation and gallstones further complicates diagnosis, highlighting the need for better screening tools. Current treatment options are limited, with surgery as the primary intervention, particularly in the early stages. Emerging liquid biopsy techniques using cfDNA offer a promising non-invasive diagnostic tool, overcoming the limitations of traditional tissue-based biopsies, providing comprehensive genomic profiling, and enabling the study of tumour evolution and resistance, which could improve GBC management in high-incidence regions ¹⁵ .

This study utilised a bead-based extraction protocol for cfDNA isolation, which offers several advantages over the silica-based method used in prior studies, such as the one by Kumari et al ¹² . Bead-based protocols provide higher efficiency, reduced contamination risk, and improved cfDNA recovery, ensuring more accurate quantification and analysis. These advancements enhance the reliability of cfDNA as a diagnostic biomarker and make it more adaptable for routine use, particularly in resource-limited settings. By addressing methodological gaps in previous research, this study contributes to refining cfDNA-based diagnostics for GBC.

Our findings align with studies by Sakamoto et al ¹¹ , Ying et al ¹⁶ , and Kumari et al ¹² , each supporting cfDNA as a potential biomarker for distinguishing GBC from benign conditions and advanced disease stages.

GBC in India shows a clear association with age, sex, and regional distribution and impacts life expectancy significantly. The disease predominantly affects older adults, with most cases occurring in individuals over 50 years of age ^2,3 . Significantly elevated levels were observed in GBC cases compared to controls. Its frequency is two to three times more common in women, likely due to hormonal factors and higher gallstone prevalence ^2,3,15 . While the findings are promising, the gender imbalance between the study and control groups should be considered, and larger studies are needed to confirm these results.

Ying et al ¹⁶ also found elevated cfDNA in GBC patients to be associated with advanced stages, jaundice, and lymph node involvement, comparable to our findings. However, their use of a chemiluminescence DNA biosensor system highlighted the high sensitivity and specificity of cfDNA as a non-invasive diagnostic marker, comparable to traditional qPCR methods.

A stratified analysis of cfDNA levels based on jaundice status revealed significantly elevated levels in jaundiced patients (1911.5±1532.2 ng/ml) compared to non-jaundiced patients (368.1±297.4 ng/ml), with a statistically significant difference (P=0.011). These findings underscored the potential confounding influence of elevated bilirubin on cfDNA levels. Incorporating jaundice-related variables is crucial for refining cfDNA-based diagnostic thresholds, particularly in GBC patients presenting with jaundice.

Wintachai et al ¹⁷ , in a study on cholangiocarcinoma, reported that cfDNA levels demonstrated superior diagnostic efficacy compared to CEA and CA19-9. Similarly, our study highlighted cfDNA’s diagnostic potential, despite weak correlations observed with traditional tumour markers such as CA-125, CA 19.9, AFP, and CEA. In contrast to these markers, cfDNA demonstrated superior diagnostic accuracy, achieving an AUC of 0.94 in ROC analysis, with high sensitivity (97.6% at 75.5 ng/ml) and specificity (100% at 188 ng/ml).

These findings underscored cfDNA’s unique diagnostic value as both a standalone marker and complementary tool to conventional tumour markers, enhancing diagnostic capabilities for GBC. While traditional markers showed limited performance, cfDNA offered insights to refine diagnostic strategies, particularly in complex scenarios where conventional markers alone could be insufficient. The integration of cfDNA with traditional biomarkers has the potential to improve diagnostic accuracy and broaden its application across patient populations, including those with benign gallbladder pathologies.

cfDNA levels were analysed as a potential prognostic marker by categorizing tumour thickness into ≤5 mm and >5 mm groups. The mean cfDNA levels were higher in the >5 mm group, but the differences were statistically insignificant (t=1.29, P=0.21). This result may reflect the complexity of cfDNA release mechanisms or the variability within the >5mm group. These findings highlight the importance of larger datasets to better characterise the relationship between tumour thickness and cfDNA levels.

Our findings established cfDNA as a versatile biomarker for GBC, offering multiple diagnostic thresholds with distinct clinical applications. Unlike previous studies ¹² , which reported a single threshold (74.37 ng/ml) with moderate diagnostic accuracy, our study provided a spectrum of diagnostic thresholds tailored to specific clinical scenarios, significantly enhancing cfDNA’s applicability.

Furthermore, while earlier research ¹² emphasised cfDNA’s integrity and global methylation status, it did not explore demographic or clinical factors influencing cfDNA levels. In contrast, our study identified significant associations between elevated cfDNA levels and factors such as gender, jaundice, and metastasis, providing a comprehensive understanding of cfDNA’s role in GBC progression and severity. This broader perspective underscores cfDNA’s diagnostic and prognostic value, particularly in resource-limited settings where traditional diagnostic methods may be unfeasible.

Additionally, our robust statistical analyses, including polynomial regression and PR curve evaluations, demonstrated superior diagnostic accuracy (AUC=0.94) and reliability. High R² values for qPCR calibration curves further validated our study’s methodological advancements. These findings collectively underscored cfDNA’s potential as a multifaceted biomarker for GBC suitable for a wide range of clinical applications.

By incorporating advanced statistical modelling, tailored diagnostic thresholds, and an in-depth analysis of clinicopathological correlations, this study builds upon and extends the foundational work of Kumari et al ¹² , offering significant advancements in the field of cfDNA-based diagnostics for GBC.

A limitation of this study is the absence of benign controls, such as chronic cholecystitis, which could better evaluate cfDNA’s specificity in differentiating malignancies from non-malignant conditions. However, the stratified analysis of cfDNA levels in jaundiced versus non-jaundiced patients provided valuable insights into bilirubin’s potential confounding effect, emphasizing the importance of expanding patient cohorts for a more comprehensive evaluation.

Sakamoto et al ¹¹ offered a unique perspective by examining cfDNA fragment lengths rather than overall levels. They observed that longer fragments were significantly reduced in advanced GBC cases with lymph node metastasis, while shorter fragments remained unchanged. This suggests that fragment length analysis could complement overall cfDNA levels in identifying advanced stages. Integrating fragment length analysis and mutation profiling into cfDNA methodologies has the potential to provide additional diagnostic and prognostic insights. Mutation profiling could uncover key genetic alterations associated with disease progression, while fragment size distribution might refine the detection of advanced or metastatic cases. Combining these approaches will likely enhance the sensitivity, specificity, and clinical utility of cfDNA-based diagnostics in GBC.

Given its non-invasive nature and high diagnostic accuracy, cfDNA testing holds promise for integration into diagnostic protocols, particularly in regions with limited access to advanced tissue-based diagnostics, thereby enabling precision oncology in under-resourced settings. To support its clinical adoption, research should evaluate the cost-effectiveness of bead-based cfDNA extraction and focus on standardising diagnostic thresholds to ensure reproducibility and scalability across diverse healthcare environments.

To the best of our knowledge, very few studies have explored cfDNA as a biomarker for GBC. This study underscores cfDNA’s potential as a non-invasive, sensitive, and specific tool for cancer screening. With the high prevalence and late-stage diagnosis of GBC, cfDNA could play a pivotal role in early detection and personalised treatment, particularly in resource-limited settings where traditional tissue-based methods are often unfeasible.

Our data revealed cfDNA levels ranging from 63.5 to 5740 ng/ml, with a mean of approximately 721 ng/ml, indicative of advanced disease stages. Elevated cfDNA levels in GBC affected individuals with lymph node involvement, jaundice, and metastasis further highlighted its utility as a marker of disease severity.

Overall, this study highlights the potential of cfDNA quantification as a non-invasive and reliable biomarker for GBC diagnosis, offering high sensitivity and specificity. By establishing multiple diagnostic thresholds tailored to distinct clinical scenarios, our findings address critical gaps in existing literature and provide a robust framework for early detection, monitoring disease progression, and personalised management of GBC. The practicality and reproducibility of cfDNA quantification make it particularly valuable in resource-limited settings where traditional diagnostic methods may be inaccessible.

In conclusion, cfDNA quantification represents a significant advancement in non-invasive cancer diagnostics, offering a scalable and effective solution for improving GBC diagnosis and, thereby, the outcome, particularly in regions with limited healthcare infrastructure.