Identification of pancreatic cancer risk factors from clinical notes using natural language processing.

OBJECTIVES: Screening for pancreatic ductal adenocarcinoma (PDAC) is considered in high-risk individuals (HRIs) with established PDAC risk factors, such as family history and germline mutations in PDAC susceptibility genes. Accurate assessment of risk factor status is provider knowledge-dependent and requires extensive manual chart review by experts. Natural Language Processing (NLP) has shown promise in automated data extraction from the electronic health record (EHR). We aimed to use NLP for automated extraction of PDAC risk factors from unstructured clinical notes in the EHR. METHODS: We first developed rule-based NLP algorithms to extract PDAC risk factors at the document-level, using an annotated corpus of 2091 clinical notes. Next, we further improved the NLP algorithms using a cohort of 1138 patients through patient-level training, validation, and testing, with comparison against a pre-specified reference standard. To minimize false-negative results we prioritized algorithm recall. RESULTS: In the test set (n = 807), the NLP algorithms achieved a recall of 0.933, precision of 0.790, and F1-score of 0.856 for family history of PDAC. For germline genetic mutations, the algorithm had a high recall of 0.851, while precision and F1-score were lower at 0.350 and 0.496 respectively. Most false positives for germline mutations resulted from erroneous recognition of tissue mutations. CONCLUSIONS: Rule-based NLP algorithms applied to unstructured clinical notes are highly sensitive for automated identification of PDAC risk factors. Further validation in a large primary-care patient population is warranted to assess real-world utility in identifying HRIs for pancreatic cancer screening.

Plasma Assay of Cell-Free Methylated DNA Markers of Colorectal Cancer: A Tumor-Agnostic Approach to Monitor Recurrence and Response to Anticancer Therapies.

BACKGROUND: Radiographic surveillance of colorectal cancer (CRC) after curative-intent therapy is costly and unreliable. Methylated DNA markers (MDMs) detected primary CRC and metastatic recurrence with high sensitivity and specificity in cross-sectional studies. This study evaluated using serial MDMs to detect recurrence and monitor the treatment response to anti-cancer therapies. METHODS: A nested case-control study was drawn from a prospective cohort of patients with CRC who completed curative-intent therapy for CRC of all stages. Plasma MDMs were assayed vis target enrichment long-probe quantitative-amplified signal assays, normalized to B3GALT6, and analyzed in combination with serum carcinoembryonic antigen to yield an MDM score. Clinical information, including treatment and radiographic measurements of the tumor burden, were longitudinally collected. RESULTS: Of the 35 patients, 18 had recurrence and 17 had no evidence of disease during the study period. The MDM score was positive in 16 out of 18 patients who recurred and only 2 of the 17 patients without recurrence. The MDM score detected recurrence in 12 patients preceding clinical or radiographic detection of recurrent CRC by a median of 106 days (range 90-232 days). CONCLUSIONS: Plasma MDMs can detect recurrent CRC prior to radiographic detection; this tumor-agnostic liquid biopsy approach may assist cancer surveillance and monitoring.

Long-Term Outcomes and Risk of Pancreatic Cancer in Intraductal Papillary Mucinous Neoplasms.

Importance: Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cysts that can give rise to pancreatic cancer (PC). Limited population data exist on their prevalence, natural history, or risk of malignant transformation (IPMN-PC). Objective: To fill knowledge gaps in epidemiology of IPMNs and associated PC risk by estimating population prevalence of IPMNs, associated PC risk, and proportion of IPMN-PC. Design, Setting, and Participants: : This retrospective cohort study was conducted in Olmsted County, Minnesota. Using the Rochester Epidemiology Project (REP), patients aged 50 years and older with abdominal computed tomography (CT) scans between 2000 and 2015 were randomly selected (CT cohort). All patients from the REP with PC between 2000 and 2019 were also selected (PC cohort). Data were analyzed from November 2021 through August 2023. Main outcomes and Measures: CIs for PC incidence estimates were calculated using exact methods with the Poisson distribution. Cox models were used to estimate age, sex, and stage-adjusted hazard ratios for time-to-event end points. Results: The CT cohort included 2114 patients (1140 females [53.9%]; mean [SD] age, 68.6 [12.1] years). IPMNs were identified in 231 patients (10.9%; 95% CI, 9.7%-12.3%), most of which were branch duct (210 branch-duct [90.9%], 16 main-duct [6.9%], and 5 mixed [2.2%] IPMNs). There were 5 Fukuoka high-risk (F-HR) IPMNs (2.2%), 39 worrisome (F-W) IPMNs (16.9%), and 187 negative (F-N) IPMNs (81.0%). After a median (IQR) follow-up of 12.0 (8.1-15.3) years, 4 patients developed PC (2 patients in F-HR and 2 patients in F-N groups). The PC incidence rate per 100 person years for F-HR IPMNs was 34.06 incidents (95% CI, 4.12-123.02 incidents) and not significantly different for patients with F-N IPMNs compared with patients without IPMNs (0.16 patients; 95% CI, 0.02-0.57 patients vs 0.11 patients; 95% CI, 0.06-0.17 patients; P = .62). The PC cohort included 320 patients (155 females [48.4%]; mean [SD] age, 72.0 [12.3] years), and 9.8% (95% CI, 7.0%-13.7%) had IPMN-PC. Compared with 284 patients with non-IPMN PC, 31 patients with IPMN-PC were older (mean [SD] age, 76.9 [9.2] vs 71.3 [12.5] years; P = .02) and more likely to undergo surgical resection (14 patients [45.2%] vs 60 patients [21.1%]; P = .003) and more-frequently had nonmetastatic PC at diagnosis (20 patients [64.5%] vs 130 patients [46.8%]; P = .047). Patients with IPMN-PC had better survival (adjusted hazard ratio, 0.62; 95% CI, 0.40-0.94; P = .03) than patients with non-IPMN PC. Conclusions and Relevance: In this study, CTs identified IPMNs in approximately 10% of patients aged 50 years or older. PC risk in patients with F-N IPMNs was low and not different compared with patients without IPMNs; approximately 10% of patients with PC had IPMN-PC, and they had better survival compared with patients with non-IPMN PC.

Plasma Methylated DNA Markers for Melanoma Surveillance.

PURPOSE: Surveillance after primary melanoma treatment aims to detect early signs of low-volume systemic disease. The current standard of care, surveillance imaging, is costly and difficult to access. We therefore sought to develop methylated DNA markers (MDMs) as promising alternatives for disease surveillance. METHODS: We used reduced representation bisulfite sequencing (RRBS) to identify MDMs in DNA samples obtained from metastatic melanoma, benign nevi, and normal skin tissues. The identified MDMs underwent validation in an independent cohort of tissue and buffy coat DNA samples. Subsequently, we tested the validated MDMs in the plasma DNA of patients with metastatic melanoma undergoing surveillance with total body imaging and compared them with cancer-free controls. To estimate the overall predictive accuracy of the MDMs, we used random forest modeling with bootstrap cross-validation. RESULTS: Forty MDMs demonstrated discrimination between melanoma cases and controls consisting of benign nevi and normal skin. Nine MDMs passing biological validation in tissue were run on 77 plasma samples from individuals with a history of metastatic melanoma, 49 of whom had evidence of disease detected by imaging at the time of blood draw, and 100 cancer-free controls. The cross-validated sensitivity of the panel for imaging-positive disease was 80% with a specificity of 100% in cancer-free controls, resulting in an overall AUC of 0.88 (95% CI, 0.81 to 0.96). The survival estimates for patients with melanoma who tested positive for the panel at 6 months and 1 year were 67% and 56%, respectively, while those who tested negative had survival rates of 100% and 92%. CONCLUSION: MDMs identified by RRBS demonstrate a high degree of concordance with imaging results in the plasma of patients with metastatic melanoma. Further prospective studies in larger intended use cohorts are needed to confirm these findings.

Assessing the capacity of methylated DNA markers of cervical squamous cell carcinoma to discriminate oropharyngeal squamous cell carcinoma in human papillomavirus mediated disease.

OBJECTIVE: Early identification of human papillomavirus associated oropharyngeal squamous cell carcinoma (HPV(+)OPSCC) is challenging and novel biomarkers are needed. We hypothesized that a panel of methylated DNA markers (MDMs) found in HPV(+) cervical squamous cell carcinoma (CSCC) will have similar discrimination in HPV(+)OPSCC tissues. MATERIALS AND METHODS: Formalin-fixed, paraffin-embedded tissues were obtained from patients with primary HPV(+)OPSCC or HPV(+)CSCC; control tissues included normal oropharynx palatine tonsil (NOP) and cervix (NCS). Using a methylation-specific polymerase chain reaction, 21 previously validated cervical MDMs were evaluated on tissue-extracted DNA. Discrimination between case and control cervical and oropharynx tissue was assessed using area under the curve (AUC). RESULTS: 34 HPV(+)OPSCC, 36 HPV(+)CSCC, 26 NOP, and 24 NCS patients met inclusion criteria. Within HPV(+)CSCC, 18/21 (86%) of MDMs achieved an AUC ≥ 0.9 and all MDMs exhibited better than chance classifications relative to control cervical tissue (all p < 0.001). In contrast, within HPV(+)OPSCC only 5/21 (24%) MDMs achieved an AUC ≥ 0.90 but 19/21 (90%) exhibited better than chance classifications relative to control tonsil tissue (all p < 0.001). Overall, 13/21 MDMs had statistically significant lower AUCs in the oropharyngeal cohort compared to the cervical cohort, and only 1 MDM exhibited a statistically significant increase in AUC. CONCLUSIONS: Previously validated MDMs exhibited robust performance in independent HPV(+)CSCC patients. However, most of these MDMs exhibited higher discrimination for HPV(+)CSCC than for HPV(+)OPSCC. This suggests that each SCC subtype requires a unique set of MDMs for optimal discrimination. Future studies are necessary to establish an MDM panel for HPV(+)OPSCC.

Methylated DNA Markers for Sporadic Colorectal and Endometrial Cancer Are Strongly Associated with Lynch Syndrome Cancers.

Lynch syndrome (LS) markedly increases risks of colorectal and endometrial cancers. Early detection biomarkers for LS cancers could reduce the needs for invasive screening and surgical prophylaxis.To validate a panel of methylated DNA markers (MDM) previously identified in sporadic colorectal cancer and endometrial cancer for discrimination of these cancers in LS.In a case-control design, previously identified MDMs for the detection of colorectal cancer and endometrial cancer were assayed by qMSP on tissue-extracted DNA. Results were normalized to ACTB values within each sample. Least absolute shrinkage and selection operator models to classify colorectal cancer and endometrial cancer were trained on sporadic cases and controls and then applied to classify colorectal cancer and endometrial cancer, in those with LS, and cross-validated.We identified colorectal cancer cases (23 with LS, 48 sporadic), colorectal controls (32 LS, 48 sporadic), endometrial cancer cases (30 LS, 48 sporadic), and endometrial controls (29 LS, 37 sporadic). A 3-MDM panel (LASS4, LRRC4, and PPP2R5C) classified LS-CRC from LS controls with an AUC of 0.92 (0.84-0.99); results were similar for sporadic colorectal cancer. A 6-MDM panel (SFMBT2, MPZ, CYTH2, DIDO1, chr10.4479, and EMX2OS) discriminated LS-EC from LS controls with an AUC of 0.92 (0.83-1.0); the AUC for sporadic endometrial cancer versus sporadic controls was nominally higher, 0.99 (0.96-1.0).MDMs previously identified in sporadic endometrial cancer and colorectal cancer discriminate between endometrial cancer and benign endometrium and colorectal cancer and benign colorectum in LS. This supports the inclusion of patients with LS within future prospective clinical trials evaluating endometrial cancer and colorectal cancer MDMs and may provide a new avenue for cancer screening or surveillance in this high-risk population. PREVENTION RELEVANCE: Lynch syndrome (LS) markedly increases risks of colorectal and endometrial cancers. Early detection biomarkers for LS cancers could reduce the needs for invasive screening and surgery. Methylated DNA markers previously identified in sporadic endometrial cancer and colorectal cancer discriminate between benign and cancer tissue in LS.

Detection of endometrial cancer using tampon-based collection and methylated DNA markers.

OBJECTIVE: Alterations in DNA methylation are early events in endometrial cancer (EC) development and may have utility in EC detection via tampon-collected vaginal fluid. METHODS: For discovery, DNA from frozen EC, benign endometrium (BE), and benign cervicovaginal (BCV) tissues underwent reduced representation bisulfite sequencing (RRBS) to identify differentially methylated regions (DMRs). Candidate DMRs were selected based on receiver operating characteristic (ROC) discrimination, methylation level fold-change between cancers and controls, and absence of background CpG methylation. Methylated DNA marker (MDM) validation was performed using qMSP on DNA from independent EC and BE FFPE tissue sets. Women ≥45 years of age with abnormal uterine bleeding (AUB) or postmenopausal bleeding (PMB) or any age with biopsy-proven EC self-collected vaginal fluid using a tampon prior to clinically indicated endometrial sampling or hysterectomy. Vaginal fluid DNA was assayed by qMSP for EC-associated MDMs. Random forest modeling analysis was performed to generate predictive probability of underlying disease; results were 500-fold in-silico cross-validated. RESULTS: Thirty-three candidate MDMs met performance criteria in tissue. For the tampon pilot, 100 EC cases were frequency matched by menopausal status and tampon collection date to 92 BE controls. A 28-MDM panel highly discriminated between EC and BE (96% (95%CI 89-99%) specificity; 76% (66-84%) sensitivity (AUC 0.88). In PBS/EDTA tampon buffer, the panel yielded 96% (95% CI 87-99%) specificity and 82% (70-91%) sensitivity (AUC 0.91). CONCLUSION: Next generation methylome sequencing, stringent filtering criteria, and independent validation yielded excellent candidate MDMs for EC. EC-associated MDMs performed with promisingly high sensitivity and specificity in tampon-collected vaginal fluid; PBS-based tampon buffer with added EDTA improved sensitivity. Larger tampon-based EC MDM testing studies are warranted.

Tissue methylated DNA markers for sporadic pancreatic cancer are strongly associated with familial and genetically predisposed pancreatic cancer.

High-risk individuals (HRIs) with familial and genetic predisposition to pancreatic ductal adenocarcinoma (PDAC) are eligible for screening. There is no accurate biomarker for detecting early-stage PDAC. We previously demonstrated that a panel of methylated DNA markers (MDMs) accurately detect sporadic PDAC. In this study we compared the distribution of MDMs in DNA extracted from tissue of PDAC cases who carry germline mutations and non-carriers with family history, with control tissue and demonstrate high discrimination like that seen in sporadic PDAC. These results provide scientific rationale for examining plasma MDMs in HRIs with the goal of developing a minimally-invasive early detection test.

Methylated DNA markers for plasma detection of ovarian cancer: Discovery, validation, and clinical feasibility.

OBJECTIVE: Aberrant DNA methylation is an early event in carcinogenesis which could be leveraged to detect ovarian cancer (OC) in plasma. METHODS: DNA from frozen OC tissues, benign fallopian tube epithelium (FTE), and buffy coats from cancer-free women underwent reduced representation bisulfite sequencing (RRBS) to identify OC MDMs. Candidate MDM selection was based on receiver operating characteristic (ROC) discrimination, methylation fold change, and low background methylation among controls. Blinded biological validation was performed using methylated specific PCR on DNA extracted from independent OC and FTE FFPE tissues. MDMs were tested using Target Enrichment Long-probe Quantitative Amplified Signal (TELQAS) assays in pre-treatment plasma from women newly diagnosed with OC and population-sampled healthy women. A random forest modeling analysis was performed to generate predictive probability of disease; results were 500-fold in silico cross-validated. RESULTS: Thirty-three MDMs showed marked methylation fold changes (10 to >1000) across all OC subtypes vs FTE. Eleven MDMs (GPRIN1, CDO1, SRC, SIM2, AGRN, FAIM2, CELF2, RIPPLY3, GYPC, CAPN2, BCAT1) were tested on plasma from 91 women with OC (73 (80%) high-grade serous (HGS)) and 91 without OC; the cross-validated 11-MDM panel highly discriminated OC from controls (96% (95% CI, 89-99%) specificity; 79% (69-87%) sensitivity, and AUC 0.91 (0.86-0.96)). Among the 5 stage I/II HGS OCs included, all were correctly identified. CONCLUSIONS: Whole methylome sequencing, stringent filtering criteria, and biological validation yielded candidate MDMs for OC that performed with high sensitivity and specificity in plasma. Larger plasma-based OC MDM studies, including testing of pre-diagnostic specimens, are warranted.

High Detection Rates of Pancreatic Cancer Across Stages by Plasma Assay of Novel Methylated DNA Markers and CA19-9.

PURPOSE: We have previously identified tissue methylated DNA markers (MDMs) associated with pancreatic ductal adenocarcinoma (PDAC). In this case-control study, we aimed to assess the diagnostic performance of plasma MDMs for PDAC. EXPERIMENTAL DESIGN: Thirteen MDMs (GRIN2D, CD1D, ZNF781, FER1L4, RYR2, CLEC11A, AK055957, LRRC4, GH05J042948, HOXA1, PRKCB, SHISA9, and NTRK3) were identified on the basis of selection criteria applied to results of prior tissue experiments and assays were optimized in plasma. Next, 340 plasma samples (170 PDAC cases and 170 controls) were assayed using target enrichment long-probe quantitative amplified signal method. Initially, 120 advanced-stage PDAC cases and 120 healthy controls were used to train a prediction algorithm at 97.5% specificity using random forest modeling. Subsequently, the locked algorithm derived from the training set was applied to an independent blinded test set of 50 early-stage PDAC cases and 50 controls. Finally, data from all 340 patients were combined, and cross-validated. RESULTS: The cross-validated area under the receiver operating characteristic curve (AUC) for the training set was 0.93 (0.89-0.96) for the MDM panel alone, 0.91 (95% confidence interval, 0.87-0.96) for carbohydrate antigen 19-9 (CA19-9) alone, and 0.99 (0.98-1) for the combined MDM-CA19-9 panel. In the test set of early-stage PDAC, the AUC for MDMs alone was 0.84 (0.76-0.92), CA19-9 alone was 0.87 (0.79-0.94), and combined MDM-CA19-9 panel was 0.90 (0.84-0.97) significantly better compared with either MDMs alone or CA19-9 alone (P = 0.0382 and 0.0490, respectively). At a preset specificity of 97.5%, the sensitivity for the combined panel in the test set was 80% (28%-99%) for stage I disease and 82% (68%-92%) for stage II disease. Using the combined datasets, the cross-validated AUC was 0.9 (0.86-0.94) for the MDM panel alone and 0.89 for CA19-9 alone (0.84-0.93) versus 0.97 (0.94-0.99) for the combined MDM-CA19-9 panel (P ≤ 0.0001). Overall, cross-validated sensitivity of MDM-CA19-9 panel was 92% (83%-98%), with an observed specificity of 92% at the preset specificity of 97.5%. CONCLUSIONS: Plasma MDMs in combination with CA19-9 detect PDAC with significantly higher accuracy compared with either biomarker individually.

Novel Methylated DNA Markers in the Surveillance of Colorectal Cancer Recurrence.

PURPOSE: We aimed to assess the concordance of colorectal cancer-associated methylated DNA markers (MDM) in primary and metastatic colorectal cancer for feasibility in detection of distantly recurrent/metastatic colorectal cancer in plasma. EXPERIMENTAL DESIGN: A panel of previously discovered colorectal cancer-associated MDMs was selected. MDMs from primary and paired metastatic colorectal cancer tissue were assayed with quantitative methylation-specific PCR. Plasma MDMs were measured blindly by target enrichment long-probe quantitative-amplified signal assays. Random forest modeling was used to derive a prediction algorithm of MDMs in archival plasma samples from primary colorectal cancer cases. This algorithm was validated in prospectively collected plasma samples from recurrent colorectal cancer cases. The accuracy of the algorithm was summarized as sensitivity, specificity, and area under the curve (AUC). RESULTS: Of the 14 selected MDMs, the concordance between primary and metastatic tissue was considered moderate or higher for 12 MDMs (86%). At a preset specificity of 95% (91%-98%), a panel of 13 MDMs, in plasma from 97 colorectal cancer cases and 200 controls, detected stage IV colorectal cancer with 100% (80%-100%) sensitivity and all stages of colorectal cancer with an AUC of 0.91 (0.87-0.95), significantly higher than carcinoembryonic antigen [AUC, 0.72 (0.65-0.79)]. This panel, in plasma from 40 cases and 60 healthy controls, detected recurrent/metastatic colorectal cancer with 90% (76%-97%) sensitivity, 90% (79%-96%) specificity, and an AUC of 0.96 (0.92-1.00). The panel was positive in 0.30 (0.19-0.43) of 60 patients with no evidence of disease in post-operative patients with colorectal cancer. CONCLUSIONS: Plasma assay of novel colorectal cancer-associated MDMs can reliably detect both primary colorectal cancer and distantly recurrent colorectal cancer with promising accuracy.

Comparison of Tissue-Based Molecular Markers in Younger versus Older Patients with Colorectal Neoplasia.

BACKGROUND: Emerging colorectal cancer trends demonstrate increased incidence and mortality in younger populations, prompting consideration of average-risk colorectal cancer screening initiation at age 45 versus 50 years. However, screening test performance characteristics in adults 45-49 years have been minimally described. To inform the biologic rationale for multi-target stool DNA (mt-sDNA) screening in younger patients, we analyzed and compared tissue levels of methylation (BMP3, NDRG4) and mutation (KRAS) markers included in the FDA-approved, mt-sDNA assay (Cologuard; Exact Sciences Corporation). METHODS: Within 40-44, 45-49, and 50-64 year age groups, archived colorectal tissue specimens were identified for 211 sporadic colorectal cancer cases, 123 advanced precancerous lesions (APLs; adenomas >1 cm, high-grade dysplasia, ≥25% villous morphology, or sessile serrated polyp; 45-49 and 50-64 age groups only), and 204 histologically normal controls. Following DNA extraction, KRAS, BMP3, and NDRG4 were quantified using QuARTS assays, relative to ACTB (reference gene). RESULTS: None of the molecular marker concentrations were significantly associated with age (P > 0.05 for all comparisons), with the exception of NDRG4 concentration in APL samples (higher in older vs. younger cases; P = 0.008). However, NDRG4 levels were also statistically higher in APL case versus normal control samples in both the 45-49 (P < 0.0001) and 50-64 (P < 0.0001) year age groups. CONCLUSIONS: Overall, these findings support the potential for earlier onset of average-risk colorectal cancer screening with the mt-sDNA assay. IMPACT: These novel data address an identified knowledge gap and strengthen the biologic basis for earlier-onset, average-risk screening with the mt-sDNA assay.

Multitarget Stool DNA Screening in Clinical Practice: High Positive Predictive Value for Colorectal Neoplasia Regardless of Exposure to Previous Colonoscopy.

OBJECTIVES: Multitarget stool DNA (MT-sDNA) testing has grown as a noninvasive screening modality for colorectal cancer (CRC), but real-world clinical data are limited in the post-FDA approval setting. The effect of previous colonoscopy on MT-sDNA performance is not known. We aimed to evaluate findings of colorectal neoplasia (CRN) at diagnostic colonoscopy in patients with positive MT-sDNA testing, stratified by patient exposure to previous colonoscopy. METHODS: We identified consecutive patients completing MT-sDNA testing over a 39-month period and reviewed the records of those with positive tests for neoplastic findings at diagnostic colonoscopy. MT-sDNA test positivity rate, adherence to diagnostic colonoscopy, and the positive predictive value (PPV) of MT-sDNA for any CRN and neoplastic subtypes were calculated. RESULTS: Of 16,469 MT-sDNA tests completed, testing returned positive in 2,326 (14.1%) patients. After exclusion of patients at increased risk for CRC, 1,801 patients remained, 1,558 (87%) of whom underwent diagnostic colonoscopy; 918 of 1,558 (59%) of these patients had undergone previous colonoscopy, whereas 640 (41%) had not. Any CRN was found in 1,046 of 1,558 patients (PPV = 67%). More neoplastic lesions were found in patients without previous colonoscopy (73%); however, the rates remained high among those who had undergone previous colonoscopy (63%, P < 0.0001). The large majority (79%) of patients had right-sided neoplasia. DISCUSSION: MT-sDNA has a high PPV for any CRN regardless of exposure to previous colonoscopy. Right-sided CRN was found at colonoscopy in most patients with positive MT-sDNA testing, representing a potential advantage over other currently available screening modalities for CRC.