ORCID Profile
0000-0001-9250-6874
Current Organisations
University of New South Wales
,
NHMRC Clinical Trials Centre
,
Garvan Institute of Medical Research
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: Cold Spring Harbor Laboratory
Date: 03-07-2022
DOI: 10.1101/2022.06.30.22277092
Abstract: Immune checkpoint blockade impedes the negative regulatory signals for T-cell response and permits more effective immune detection and eradication of cancer cells. This single-arm phase II clinical trial (ACTRN12616001019493) within the Molecular Screening and Therapeutics (MoST) program evaluates the clinical activity and safety of combination immunotherapy with durvalumab and tremelimumab in patients with advanced cancers, prioritsing rare cancers ( per 100,000 annual incidence) and patients having failed standard treatments for their cancer type. Eligible patients were determined by the molecular tumour board based on the absence of actionable genomic findings (n=64) and biomarker enriched (n=48) at screening. Patients received durvalumab 1500 mg and tremelimumab 75 mg every four weeks for 4 cycles, followed by durvalumab alone for another 9 cycles. The primary endpoint was progression-free survival at 6 months (PFS6) and secondary endpoints included objective response, time to progression (TTP) on trial to TTP on prior therapy (TTP2/TTP1 .3), overall survival and treatment tolerability. Between December 2016 and 2019, 112 patients were enrolled on the study. There was a female predominance (55%), most had an ECOG performance status of 0 (66%), aged years (75%), with rare cancers (84%). The PFS6 rate was 32% (95% CI 23 to 40%) 16 of 112(14%) achieved an objective response TTP2/TTP1 .3 for 22 of 63 (35%) patients with an evaluable ratio median overall survival 11.9 months (95% CI 11.0 to 14.8), and there were no new safety concerns. High tumour cell PD-L1 correlated with improved PFS and OS and TMB with PFS alone. More PD-1 + CD4 + T-cells and circulating follicular T-helper (cTfh) cells at baseline were strongly associated with better PFS and OS. Durvalumab plus tremelimumab demonstrated a signal of clinical activity in treatment-refractory patients with rare cancers. A PFS6 of 32% and 35% of patients achieving a TTP2/TTP1 .3 suggests an improved disease trajectory on trial. Translational correlates provided insights into biological associations with clinical outcomes across tumour types.
Publisher: Springer Science and Business Media LLC
Date: 31-07-2017
DOI: 10.1038/S41598-017-07111-0
Abstract: Vast amounts of clinically relevant text-based variables lie undiscovered and unexploited in electronic medical records (EMR). To exploit this untapped resource, and thus facilitate the discovery of informative covariates from unstructured clinical narratives, we have built a novel computational pipeline termed T ext-based E xploratory P attern A nalyser for P rognosticator and A ssociator discovery (TEPAPA). This pipeline combines semantic-free natural language processing (NLP), regular expression induction, and statistical association testing to identify conserved text patterns associated with outcome variables of clinical interest. When we applied TEPAPA to a cohort of head and neck squamous cell carcinoma patients, plausible concepts known to be correlated with human papilloma virus (HPV) status were identified from the EMR text, including site of primary disease, tumour stage, pathologic characteristics, and treatment modalities. Similarly, correlates of other variables (including gender, nodal status, recurrent disease, smoking and alcohol status) were also reliably recovered. Using highly-associated patterns as covariates, a patient’s HPV status was classifiable using a bootstrap analysis with a mean area under the ROC curve of 0.861, suggesting its predictive utility in supporting EMR-based phenotyping tasks. These data support using this integrative approach to efficiently identify disease-associated factors from unstructured EMR narratives, and thus to efficiently generate testable hypotheses.
Publisher: Cold Spring Harbor Laboratory
Date: 22-12-2020
DOI: 10.1101/2020.12.18.20248521
Abstract: While several key resources exist that interpret therapeutic significance of genomic alterations in cancer, many regional real-world issues limit access to drugs. There is a need for a pragmatic, evidence-based, context-adapted tool to guide clinical management based on molecular biomarkers. A compendium of approved and experimental therapies with associated biomarkers was built following a survey of drug regulatory databases, existing knowledge bases, and published literature. Each biomarker-disease-therapy triplet was then categorized using a tiering system reflective of key therapeutic considerations: approved and reimbursed standard-of-care therapies with respect to a jurisdiction (Tier 1), evidence of efficacy or approval in another jurisdiction (Tier 2), evidence of antitumour activity (Tier 3), and plausible biological rationale (Tier 4). Two resistance categories were defined: lack of efficacy (Tier R1), and lack of antitumor activity (Tier R2). Following comprehensive literature review and appraisal, we developed a curated knowledge base focused on drugs relevant and accessible in the Australian healthcare system (TOPOGRAPH: Therapy Oriented Precision Oncology Guidelines for Recommending Anticancer Pharmaceuticals). As of November 2020, TOPOGRAPH comprised 2810 biomarker-disease-therapy triplets in 989 expert-appraised entries, including 373 therapies, 199 predictive biomarkers, and 106 cancer types. In the 345 biomarker-linked therapies catalogued, 84 (24%) and 65 (19%) therapies in contexts of different cancer types have Tier 1 and 2 designations respectively, while 271 (79%) therapies were supported by preclinical studies, early clinical trials, retrospective studies, or case series (Tiers 3 and 4). A total of 119 of 373 (33%) therapies associated with biomarkers of resistance were also catalogued. A clinical algorithm was also developed to support therapeutic decision-making using predictive biomarkers. This resource is accessible online at topograph.info/ . TOPOGRAPH is intended to support oncologists with context-appropriate clinical decision-making– optimising selection and accessibility of the most appropriate targeted therapy for any given genomic biomarker. Our approach can be readily adapted to build jurisdiction-specific resources to standardise decision-making in precision oncology.
Publisher: Springer Science and Business Media LLC
Date: 23-06-2021
DOI: 10.1038/S41698-021-00194-Z
Abstract: While several resources exist that interpret therapeutic significance of genomic alterations in cancer, many regional real-world issues limit access to drugs. There is a need for a pragmatic, evidence-based, context-adapted tool to guide clinical management based on molecular biomarkers. To this end, we have structured a compendium of approved and experimental therapies with associated biomarkers following a survey of drug regulatory databases, existing knowledge bases, and published literature. Each biomarker-disease-therapy triplet was categorised using a tiering system reflective of key therapeutic considerations: approved and reimbursed therapies with respect to a jurisdiction (Tier 1), evidence of efficacy or approval in another jurisdiction (Tier 2), evidence of antitumour activity (Tier 3), and plausible biological rationale (Tier 4). Two resistance categories were defined: lack of efficacy (Tier R1) or antitumor activity (Tier R2). Based on this framework, we curated a digital resource focused on drugs relevant in the Australian healthcare system (TOPOGRAPH: Therapy Oriented Precision Oncology Guidelines for Recommending Anticancer Pharmaceuticals). As of November 2020, TOPOGRAPH comprised 2810 biomarker-disease-therapy triplets in 989 expert-appraised entries, including 373 therapies, 199 biomarkers, and 106 cancer types. In the 345 therapies catalogued, 84 (24%) and 65 (19%) were designated Tiers 1 and 2, respectively, while 271 (79%) therapies were supported by preclinical studies, early clinical trials, retrospective studies, or case series (Tiers 3 and 4). A companion algorithm was also developed to support rational, context-appropriate treatment selection informed by molecular biomarkers. This framework can be readily adapted to build similar resources in other jurisdictions to support therapeutic decision-making.
Publisher: American Society of Clinical Oncology (ASCO)
Date: 10-2022
DOI: 10.1200/CCI.22.00064
Abstract: Predicting short-term mortality in patients with advanced cancer remains challenging. Whether digitalized clinical text can be used to build models to enhance survival prediction in this population is unclear. We conducted a single-centered retrospective cohort study in patients with advanced solid tumors. Clinical correspondence authored by oncologists at the first patient encounter was extracted from the electronic medical records. Machine learning (ML) models were trained using narratives from the derivation cohort, before being tested on a temporal validation cohort at the same site. Performance was benchmarked against Eastern Cooperative Oncology Group performance status (PS), comparing ML models alone (comparison 1) or in combination with PS (comparison 2), assessed by areas under receiver operating characteristic curves (AUCs) for predicting vital status at 11 time points from 2 to 52 weeks. ML models were built on the derivation cohort (4,791 patients from 2001 to April 2017) and tested on the validation cohort of 726 patients (May 2017-June 2019). In 441 patients (61%) where clinical narratives were available and PS was documented, ML models outperformed the predictivity of PS (mean AUC improvement, 0.039, P .001, comparison 1). Inclusion of both clinical text and PS in ML models resulted in further improvement in prediction accuracy over PS with a mean AUC improvement of 0.050 ( P .001, comparison 2) the AUC was 0.80 at all assessed time points for models incorporating clinical text. Exploratory analysis of oncologist's narratives revealed recurring descriptors correlating with survival, including referral patterns, mobility, physical functions, and concomitant medications. Applying ML to oncologists' narratives with or without including patient's PS significantly improved survival prediction to 12 months, suggesting the utility of clinical text in building prognostic support tools.
Publisher: Cold Spring Harbor Laboratory
Date: 30-10-2020
DOI: 10.1101/2020.10.28.20214627
Abstract: Electronic medical records (EMR) represent a rich informatics resource that remains largely unexploited for improving healthcare outcomes. Here we report a systematic text mining analysis of EMR correspondence for 4791 cancer patients treated between 2001 and 2017. Meaningful groups of text descriptors correlating with poor survival outcomes were systematically identified, and applying machine learning analysis to clinical text accurately predicted cancer patient survival at selected timepoints up to 12 months. In a validation cohort of 726 patients, inclusion of EMR descriptors to machine learning models outperformed the predictivity of conventional clinical symptom scores by 4.9% ( p = 0.001). These results prove that labour-intensive EMR data collection can be repurposed to add clinical value. Extension of this approach to a broader spectrum of digital health data should transform the real-time utility of such latent informatics resources, enabling healthcare systems to be more adaptive and responsive to patient circumstances.
No related grants have been discovered for Frank Lin.