Off-label targeted therapy (TT) use in recurrent/metastatic NSCLC.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. e18688-e18688
Author(s):  
Victoria Wang ◽  
Chenming Cui ◽  
Lei Yang ◽  
Gerald Li ◽  
Alexa Betzig Schrock ◽  
...  
Keyword(s):  
Targeted Therapy ◽  
Clinical Efficacy ◽  
Real World ◽  
Survival Advantage ◽  
Genomic Alteration ◽  
Genomic Profiling ◽  
Off Label ◽  
Number Of Patients ◽  
Tier System ◽  
Line Of Therapy

e18688 Background: Comprehensive genomic profiling (CGP) is an established diagnostic approach to select patients for TT. As CGP gains wide adoption, an increasing number of patients are found to harbor driver mutations for which no approved TT is available. This is often addressed through use of matched TKI and mAb approved for other mutations or anatomic sites. In this work, we examined the clinical efficacy of off-label TT in R/M NSCLC. Methods: Using a de-identified NSCLC clinico-genomic database (CGDB), we identified 6590 NSCLC patients who underwent Foundation Medicine CGP, of whom 17.8% harbored an actionable genomic alteration (GA) for which an FDA-approved TT was available and 2% (133) whose GA (MET ex-14, uncommon EGFRm, EGFR ex20ins, HER2 amp/mut, RET fusion, BRAF class 2/3) lacked an FDA-approved TT (62 in 1L and 71 in ≥2L ) who received matched off-label TT. ESMO Scale for Clinical Actionability (ESCAT) was used to grade levels of evidence. For patients who progressed on initial chemotherapy (range 2 – 9 lines, median 3), we calculated clinical efficacy using the ratio of real world PFS on targeted therapy (rw-PFS2) to rw-PFS on the last prior line of therapy (rw-PFS1) and used a cut-off of PFS2/PFS1 > 1.3 to determine off-label drug efficacy. Results: Of the 133 patients reviewed, 72 were classified as ESCAT level IB (uncommon EGFRm, MET-ex14), 45 IIB (HER2m/amp, EGFR ex-20ins), 7 IC (RET fusions). PFS varied significantly by mutation and line of therapy (table 1) with uncommon EGFRm+ and MET-ex14 exhibiting best response while EGFR ex20 ins, BRAF class 2/3 and HER2 amp fared significantly worse. 55.8% of the patients (39 of 71) reached a PFS2/PFS1 ratio > 1.3 (two-sided 95% CI, 45.3 % – 68.7 %), ranging from 93% in uncommon EGFRm+ down to 20% in HER2 amp and 44% in ex20ins. Conclusions: We provide real-world evidence to assess off-label TT in NSCLC. Clinical benefit derived from off-label TT is unevenly distributed across various mutations with most survival advantage accruing to specific mutations (MET-ex14 and uncommon EGFRm) at the expense of others (HER2 amp). Survival advantage was highly influenced by two factors: A) the timing of CGP with the earlier recipients of genomic profiling achieving better outcome, B) the identity of the driver mutation, highlighting the role of clinical actionability tier system to define level of evidence supporting such intervention.[Table: see text]

Precision oncology experience at a tertiary care center.

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. e18118-e18118
Author(s):  
Meena Sadaps ◽  
Pauline Funchain ◽  
Petros Grivas ◽  
Bassam N. Estfan ◽  
Jame Abraham ◽  
...  
Keyword(s):  
Clinical Trials ◽  
Targeted Therapy ◽  
Real World ◽  
Care Center ◽  
Tertiary Care ◽  
Cleveland Clinic ◽  
Tumor Board ◽  
Genomic Testing ◽  
Precision Oncology ◽  
Genomic Profiling

e18118 Background: Precision oncology – use of tumor genomic profiling to guide therapies – is widely discussed but with limited real-world data. We have previously reported our prospective study on feasibility and clinical utility of routine somatic genomic testing of solid tumors [ J Natl Cancer Inst. 2015; 108(3)], and here we report our longitudinal experience, focusing on therapeutic impact. Methods: Records were reviewed for consecutive adult patients seen at Cleveland Clinic for a solid tumor malignancy without known curative options where tumor genomic profiling was ordered using FoundationOne™ (Cambridge, MA). Results were discussed at the Cleveland Clinic Genomics Tumor Board, and therapeutic recommendations were conveyed to the primary oncologist. Data for this cohort study approved by the Cleveland Clinic IRB included subsequent therapies and clinical outcomes. Results: From 2013 to 2016, 330 patients had tumor genomic testing ordered. Median age was 61 years (range, 24-94); 170 (51.5%) were female; 289 (87.6%) were Caucasian. Colorectal (21.5%), breast (17%), lung (16.1%), and pancreatobiliary (11.5%) cancers were the most common diagnoses. In 300 resulted cases, a median of 4 (0-20) alterations per specimen were noted; the most commonly altered genes were TP53 (n = 174), KRAS (n = 75), APC (n = 65), CDKN2A/B (n = 49), and PIK3CA/ PIK3R (n = 46). A specific therapy targeting an actionable alteration was recommended in 51% (153/300) of patients, and 11.7% (n = 35) received such therapy: 14 on clinical trials, 5 on-label, and 16 off-label. Most common targets for therapy were PIK3CA/PIK3R/PTEN (n = 7), HER2 (6), BRAF (3), EGFR (3), and ALK, FLT3, NTRK1 and RET (2 each). At last follow-up, of 35 patients receiving targeted therapy, best responses were: complete response (n = 1, 2.9%), partial response (n = 5, 14.3%), stable disease (n = 14, 40%), progressive disease (n = 11, 31.4%); data not available for 4 patients. Non-availability of clinical trials was a common reason for non-receipt of targeted therapy. Conclusions: Tumor genomic profiling influenced treatment in 11.7% of patients in this cohort, and 57% of those receiving targeted therapy experienced clinical benefit. These data can help guide real-world discussions of precision oncology.

The genomic landscape of gene fusions across solid tumors and clinical outcome of targeted therapies: A real-world retrospective analysis.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. 3132-3132
Author(s):  
Yumeng Zhang ◽  
Bindiya G Patel ◽  
Todd C Knepper ◽  
Dung-Tsa Chen ◽  
Jhanelle Elaine Gray ◽  
...  
Keyword(s):  
Targeted Therapy ◽  
Solid Tumors ◽  
Solid Tumor ◽  
Real World ◽  
Targeted Therapies ◽  
Gene Fusion ◽  
Treatment Modalities ◽  
Gene Fusions ◽  
Sequencing Data ◽  
Off Label

3132 Background: Oncogenic gene fusions can be observed across numerous solid tumor types with therapies targeting fusion events emerging as important treatment modalities. The occurrence of rare fusion events and response to targeted therapy inclusive of off-label drug use has not been fully elucidated. We describe a real-world (RW) landscape of gene fusions in solid tumors and treatment outcomes of targeted therapies. Methods: Patients with solid tumors harboring a gene fusion or rearrangement were retrospectively identified through review of a clinical molecular database housing sequencing data on 6,800 patients from a single-center between 1/1/2015 – 12/31/2019. Patients who received targeted therapy for gene fusions were divided into three arms: off-label, on-label, and clinical trial use. Clinical characteristics were summarized using descriptive statistics. Overall survival (OS) and Progression free survival (PFS) between the three arms were compared using the Kaplan-Meier estimates. Results: A total of 336 (4.9%) patients had a fusion positive solid tumor with 197 (2.9%) having a fusion event predicted to be oncogenic and could be targeted with a drug. Thirty different cancer types had targetable fusions with the three most common types being lung adenocarcinoma (41%), glioblastoma (10.2%), and melanoma (7.1%). The most common observed targetable fusions included ALK (21.3%), RET (11.7%), ROS1 (9.1%), and FGFR2 (8.1%). A total of 71 patients received targeted therapy; 37 (52%) received therapies on-label, 20 (28%) off-label, and 14 (20%) on-trial (Table). The median PFS was 4 months for off-label, 16 months for on-label, and 9 months for on trial-therapy (p=0.02). The median OS was 8 months for off-label, 51 months for on-label, and 11 months for on-trial therapy (p=0.001). Seven out of twenty patients (35%) in the off-label group had PFS for at least 6 months. Three patients had a response for more than one year. However, higher toxicity related discontinuation rate was observed (30%, 8%, 7% for off-label, on-label, and on-trial, p=0.03). Conclusions: Off-label targeted therapy had shorter PFS and OS when compared to on-label therapy. However, 35% patients in the off-label group had at least 6 months PFS. Off-label therapy remained a valuable option for patients who were not candidate for clinical trials or with rare fusions. Further studies are needed to determine which patients are most likely to benefit from targeting gene fusion events.[Table: see text]

scholarly journals O32 Serious infection with tocilizumab compared to TNF-inhibitors and other bDMARDS in rheumatoid arthritis patients: does line of therapy matter?

Rheumatology ◽  
2021 ◽  
Vol 60 (Supplement_1) ◽  
Author(s):  
Kim Lauper ◽  
Lianne Kearsley-Fleet ◽  
Rebecca Davies ◽  
Kath Watson ◽  
Mark Lunt ◽  
...  
Keyword(s):  
Rheumatoid Arthritis ◽  
Real World ◽  
Disease Duration ◽  
Tnf Inhibitors ◽  
World Population ◽  
Research Support ◽  
Serious Infection ◽  
Number Of Patients ◽  
Serious Infections ◽  
Line Of Therapy

Abstract Background/Aims  In the real-world, tocilizumab is prescribed to a population of patients different from those prescribed TNF-inhibitors, often older with longer disease duration, worse functional status and more previous b- or tsDMARDs. The aim of this study was to evaluate if and how the risk of serious infection on tocilizumab and other bDMARDs differs when stratifying by line of therapy in a real-world population of rheumatoid arthritis patients. Methods  We included patients registered in the BSRBR-RA treated with tocilizumab, etanercept, adalimumab, infliximab, certolizumab, abatacept or rituximab, including biosimilars. Primary outcome was the occurrence of a serious infection (defined as infection requiring hospitalisation, intravenous antibiotics or resulting in death). Primary covariate of interest was line of therapy (from first to fifth line of therapy). Every change to another b- or tsDMARD was considered a new line of therapy, but not a change between a bio-original and a biosimilar. Hazard ratios (HR) of serious infections were estimated using an inverse probability weighted Cox regression, based on a propensity score including baseline patient and disease characteristics, and adjusting for time in study (see table). The reference group was etanercept, which included the highest number of patients. Treatment exposure was analysed without and with stratification by line of therapy. Results  A total of 33,916 treatment courses were included (Table) contributing to 62,532 years of follow-up. Compared to etanercept, participants starting abatacept, tocilizumab and rituximab were older, had more previous bDMARDs, longer disease duration and more comorbidities. The crude HR of serious infections were higher with infliximab and adalimumab, lower with certolizumab and rituximab, and not significantly different for abatacept and tocilizumab compared to etanercept. After adjustment, HR of serious infections were higher with tocilizumab, adalimumab and infliximab. However, when stratified by line of therapy, HR were no longer significantly different compared to etanercept for tocilizumab, adalimumab and infliximab for most lines of therapy. Conclusion  Whilst initially there appears to be a difference in rates of serious infections between biologic therapies, line of therapy may be a confounding factor when comparing the risk of serious infections between bDMARDs. Disclosure  K. Lauper: Honoraria; Gilead-Galapagos. Grants/research support; AbbVie. Other; AbbVie, Pfizer. L. Kearsley-Fleet: None. R. Davies: None. K. Watson: None. M. Lunt: None. K.L. Hyrich: Honoraria; AbbVie. Grants/research support; Pfizer, BMS.

Next-generation sequencing (NGS) in metastatic breast cancer (mBC) patients: Translation from sequence data into clinical practice.

2015 ◽  
Vol 33 (28_suppl) ◽  
pp. 133-133 ◽  
Author(s):  
Raphael Brandao Moreira ◽  
Renata Peixoto ◽  
Tercia Tarciane Soares de Sousa ◽  
Marcelo Cruz ◽  
Fernando Maluf ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Targeted Therapy ◽  
Sequence Data ◽  
Objective Response ◽  
Metastatic Breast ◽  
Genomic Alteration ◽  
Histologic Subtype ◽  
Mutational Profiling ◽  
Number Of Patients ◽  
Her 2

133 Background: A considerable number of patients with mBC progress after exhausting all approved standard therapies but maintain an adequate performance status and could be candidates for further treatment. We aim at reviewing our experience with the use of a NGS platform in refractory mBC and its clinical utility. Methods: We retrospectively reviewed demographics, NGS results, and the suggested therapies received by patients undergoing NGS (Foundation Medicine, Cambridge, MA, USA): exonic sequencing of 315 genes and selective intronic sequencing from 28 genes for refractory mBC. Co-primary endpoints were the percentage of patients with targeted therapy options uncovered by mutational profiling and the percentage of them who received genotype-directed therapy. Results: Samples from 20 patients were tested. Histologic subtype consisted of triple negative (50%), HER 2 +, hormonal receptor [HR] - 15%, HR + and HER 2 - (35%). A targetable genomic alteration was identified in 14 (70%) patients, most frequently in TP53 (11 [55%]), PIK3CA (8 [40%]), FGFR1 (4 [20%]), CCND1 (3 [15%]), PTEN (3[15%]) and BRCA2 (2[10%]). Therapy could be personalized in 9 (45%) of 14 patients. Of the 9 patients who were received targeted therapy, 7 (77%) had an objective response and had stable disease for more than 3 months. Conclusions: Mutational profiling using a targeted NGS panel identified potentially actionable alterations in the majority of advanced breast cancer patients. The assay provided clinical benefit in 35% of the patients.

Biomarker testing patterns and use of targeted therapy in Medicare Fee-for-Service (FFS) beneficiaries newly diagnosed with metastatic non-small cell lung cancer (mNSCLC).

2020 ◽  
Vol 38 (29_suppl) ◽  
pp. 278-278
Author(s):  
Anik Patel ◽  
Aurelien Jamotte ◽  
Tara Matsuda ◽  
Partha Manas Das ◽  
Ehab Elkhouly ◽  
...  
Keyword(s):  
Targeted Therapy ◽  
Targeted Therapies ◽  
Single Gene ◽  
Fee For Service ◽  
Newly Diagnosed ◽  
Gene Testing ◽  
Number Of Patients ◽  
Biomarker Testing ◽  
Line Of Therapy ◽  
Test Result

278 Background: The number of targeted therapies approved for treatment of mNSCLC has increased over the past 5 years. Strategies to identify eligible patients with actionable mutations for targeted therapy include simultaneous testing of ≥ 2 genes via next generation sequencing (NGS) or multiple simultaneous gene testing (MSGT) and sequential single gene testing (SSGT). Current clinical practice guidelines strongly recommend broad molecular profiling in all patients for the simultaneous assessment of multiple genes, including EGFR, ALK and ROS1, that may have potential roles in cancer development. Limited real-world (RW) evidence is available describing the uptake of these strategies and receipt of targeted therapy. Methods: Medicare beneficiaries age 65 years or older, newly diagnosed with mNSCLC and tested for mutations of interest in mNSCLC (ALK, EGFR, ROS1, BRAF, HER2, KRAS, MET, NTRK, RET) from July 2014 - June 2018 were identified using Medicare FFS claims (100% sample) linked to biomarker results in PROGNOS NSCLC Explorer. Patients were followed from date of first metastatic diagnosis and stratified by line of therapy, testing strategy, and year of mNSCLC diagnosis. Those testing positive for an actionable biomarker were identified and then segmented by timing of receipt of a subsequent targeted therapy. Results: 12,272 beneficiaries met inclusion criteria: median age: 75 years, 51% were female, 86% white. Among mNSCLC patients with at least one biomarker test result, EGFR and ALK mutation status were the most commonly tested and reported in 85% and 63% respectively. Overall, 1540 (12.5%) tested positive for EGFR, ALK or ROS1. The relative use of NGS or MSGT vs. SSGT for biomarker testing increased over time, from 63% in 2014 to 80% in 2018. During this period, 789 patients were identified as having at least one positive biomarker test result prior to initiating 1L therapy: 635 were identified via NGS or MSGT while 154 were identified via SSGT. Despite a positive test for mutations of interest, only 292 patients received a targeted drug at 1L. Conclusions: This RW study of mNSCLC patients demonstrates an increasing trend to test patients for multiple biomarkers at once via NGS or other MSGT methods. The number of patients receiving appropriate targeted therapies was low, suggesting the need to address the barriers to administration of guideline-recommended therapy.

scholarly journals Adverse reactions of targeted therapy in cancer patients: a retrospective study of hospital medical data in China

BMC Cancer ◽  
2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Ruofei Du ◽  
Xin Wang ◽  
Lixia Ma ◽  
Leon M. Larcher ◽  
Han Tang ◽  
...  
Keyword(s):  
Targeted Therapy ◽  
Cancer Patients ◽  
Adverse Reactions ◽  
Cox Regression ◽  
Target Therapy ◽  
Skin Damage ◽  
Data Set ◽  
Patients With Cancer ◽  
Number Of Patients ◽  
Significant Difference

Abstract Background The adverse reactions (ADRs) of targeted therapy were closely associated with treatment response, clinical outcome, quality of life (QoL) of patients with cancer. However, few studies presented the correlation between ADRs of targeted therapy and treatment effects among cancer patients. This study was to explore the characteristics of ADRs with targeted therapy and the prognosis of cancer patients based on the clinical data. Methods A retrospective secondary data analysis was conducted within an ADR data set including 2703 patients with targeted therapy from three Henan medical centers of China between January 2018 and December 2019. The significance was evaluated with chi-square test between groups with or without ADRs. Univariate and multivariate logistic regression with backward stepwise method were applied to assess the difference of pathological characteristics in patients with cancer. Using the univariate Cox regression method, the actuarial probability of overall survival was performed to compare the clinical outcomes between these two groups. Results A total of 485 patients were enrolled in this study. Of all patients, 61.0% (n = 296) occurred ADRs including skin damage, fatigue, mucosal damage, hypertension and gastrointestinal discomfort as the top 5 complications during the target therapy. And 62.1% of ADRs were mild to moderate, more than half of the ADRs occurred within one month, 68.6% ADRs lasted more than one month. Older patients (P = 0.022) and patients with lower education level (P = 0.036), more than 2 comorbidities (P = 0.021), longer medication time (P = 0.022), drug combination (P = 0.033) and intravenous administration (P = 0.019) were more likely to have ADRs. Those with ADRs were more likely to stop taking (P = 0.000), change (P = 0.000), adjust (P = 0.000), or not take the medicine on time (P = 0.000). The number of patients with recurrence (P = 0.000) and metastasis (P = 0.006) were statistically significant difference between ADRs and non-ADRs group. And the patients were significantly poor prognosis in ADRs groups compared with non-ADRs group. Conclusion The high incidence of ADRs would affect the treatment and prognosis of patients with cancer. We should pay more attention to these ADRs and develop effective management strategies.

Endovascular aortic repair with EndoAnchors demonstrate good mid-term outcomes in physician-initiated multicenter analysis—The PERU registry

Vascular ◽  
2021 ◽  
pp. 170853812199259
Author(s):  
Andrés Reyes Valdivia ◽  
Arindam Chaudhuri ◽  
Ross Milner ◽  
Giovanni Pratesi ◽  
Michel MPJ Reijnen ◽  
...  
Keyword(s):  
Real World ◽  
Endovascular Aneurysm Repair ◽  
Abdominal Ultrasound ◽  
Interquartile Range ◽  
Technical Success ◽  
Number Of Patients ◽  
Hostile Neck ◽  
Aneurysm Repair ◽  
Related Mortality

Objectives We aim to describe real-world outcomes from multicenter data about the efficacy of adjunct Heli-FX EndoAnchor usage in preventing or repairing failures during infrarenal endovascular aneurysm repair (EVAR), so-called EndoSutured-aneurysm-repair (ESAR). Methods The current study has been assigned an identifier (NCT04100499) at the US National Library of Medicine ( https://ClinicalTrials.gov ). It is an observational retrospective study of prospectively collected data from seven vascular surgery departments between June 2010 and December 2019. Patients included in the ANCHOR registry were excluded from this analysis. The decision for the use of EndoAnchors was made by the treating surgeon or multidisciplinary aortic committee according to each center’s practice. Follow-up imaging was scheduled according to each center’s protocol, which necessarily included either abdominal ultrasound or radiography or computed tomographic scan imaging. The main outcomes analyzed were technical success, freedom from type Ia endoleaks (IaEL), all-cause and aneurysm-related mortality, and sac variation and trends evaluated for those with at least six months imaging follow-up. Results Two hundred and seventy-five patients underwent ESAR in participating centers during the study period. After exclusions, 221 patients (184 males, 37 females, mean age 75 ± 8.3 years) were finally included for analysis. Median follow-up for the cohort was 27 (interquartile range 12–48) months. A median 6 (interquartile range 3) EndoAnchors were deployed at ESAR, 175 (79%) procedures were primary and 46 (21%) revision cases, 40 associated with type IaEL. Technical success at operation (initial), 30-day, and overall success were 89, 95.5, and 96.8%, respectively; the 30-day success was higher due to those with subsequent spontaneous proximal endoleak seal. At two years, freedom from type IaEL was 94% for the whole series; 96% and 86% for the primary and revision groups, respectively; whereas freedom from all-cause mortality, aneurysm-related mortality, and reintervention was 89%, 98%, and 87%, respectively. Sac evolution pre-ESAR was 66 ± 15.1 vs. post ESAR 61 ± 17.5 (p < 0.001) and for 180 patients with at least six-month follow-up, 92.2% of them being in a stable (51%) or regression (41%) situation. Conclusions This real-world registry demonstrates that adjunct EndoAnchor usage at EVAR achieves high rates of freedom from type IaEL at mid-term including in a high number of patients with hostile neck anatomy, with positive trends in sac-size evolution. Further data with longer follow-up may help to establish EndoAnchor usage as a routine adjunct to EVAR, especially in hostile necks.

scholarly journals Clinical efficacy of chloroquine derivatives in COVID-19 infection: comparative meta-analysis between the big data and the real world

2020 ◽  
Vol 38 ◽  
pp. 100709 ◽  
Author(s):  
M. Million ◽  
P. Gautret ◽  
P. Colson ◽  
Y. Roussel ◽  
G. Dubourg ◽  
...  
Keyword(s):  
Big Data ◽  
Clinical Efficacy ◽  
Real World ◽  
Meta Analysis ◽  
The Real

Real‐world off‐label use of icatibant for acute management of non‐hereditary angioedema

2021 ◽  
Vol 51 (3) ◽  
pp. 419-423
Author(s):  
Thanh‐Thao (Adriana) Le ◽  
William Smith ◽  
Pravin Hissaria
Keyword(s):  
Real World ◽  
Hereditary Angioedema ◽  
Acute Management ◽  
Off Label Use ◽  
Off Label ◽  
Label Use

scholarly journals Generating real-world evidence from unstructured clinical notes to examine clinical utility of genetic tests: use case in BRCAness

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Yiqing Zhao ◽  
Saravut J. Weroha ◽  
Ellen L. Goode ◽  
Hongfang Liu ◽  
Chen Wang
Keyword(s):  
Targeted Therapy ◽  
Data Quality ◽  
Real World ◽  
Genetic Information ◽  
Genetic Data ◽  
Real World Data ◽  
Rule Based ◽  
Clinical Notes ◽  
Real World Evidence ◽  
F Measure

Abstract Background Next-generation sequencing provides comprehensive information about individuals’ genetic makeup and is commonplace in oncology clinical practice. However, the utility of genetic information in the clinical decision-making process has not been examined extensively from a real-world, data-driven perspective. Through mining real-world data (RWD) from clinical notes, we could extract patients’ genetic information and further associate treatment decisions with genetic information. Methods We proposed a real-world evidence (RWE) study framework that incorporates context-based natural language processing (NLP) methods and data quality examination before final association analysis. The framework was demonstrated in a Foundation-tested women cancer cohort (N = 196). Upon retrieval of patients’ genetic information using NLP system, we assessed the completeness of genetic data captured in unstructured clinical notes according to a genetic data-model. We examined the distribution of different topics regarding BRCA1/2 throughout patients’ treatment process, and then analyzed the association between BRCA1/2 mutation status and the discussion/prescription of targeted therapy. Results We identified seven topics in the clinical context of genetic mentions including: Information, Evaluation, Insurance, Order, Negative, Positive, and Variants of unknown significance. Our rule-based system achieved a precision of 0.87, recall of 0.93 and F-measure of 0.91. Our machine learning system achieved a precision of 0.901, recall of 0.899 and F-measure of 0.9 for four-topic classification and a precision of 0.833, recall of 0.823 and F-measure of 0.82 for seven-topic classification. We found in result-containing sentences, the capture of BRCA1/2 mutation information was 75%, but detailed variant information (e.g. variant types) is largely missing. Using cleaned RWD, significant associations were found between BRCA1/2 positive mutation and targeted therapies. Conclusions In conclusion, we demonstrated a framework to generate RWE using RWD from different clinical sources. Rule-based NLP system achieved the best performance for resolving contextual variability when extracting RWD from unstructured clinical notes. Data quality issues such as incompleteness and discrepancies exist thus manual data cleaning is needed before further analysis can be performed. Finally, we were able to use cleaned RWD to evaluate the real-world utility of genetic information to initiate a prescription of targeted therapy.

Sign in / Sign up

Export Citation Format

Share Document