scholarly journals Are Machines-learning Methods More Efficient than Humans in Triaging Literature for Systematic Reviews?

2021 ◽  
Author(s):  
Seye Abogunrin ◽  
Luisa Queiros ◽  
Mateusz Bednarski ◽  
Marc Sumner ◽  
David Baehrens ◽  
...  

Systematic literature reviews provide rigorous assessments of clinical, cost-effectiveness, and humanistic data. Accordingly, there is a growing trend worldwide among healthcare agencies and decision-makers to require them in order to make informed decisions. Because these reviews are labor-intensive and time consuming, we applied advanced analytic methods (AAM) to determine if machine learning methods could classify abstracts as well as humans. Literature searches were run for metastatic non-small cell lung cancer treatments (mNSCLC) and metastatic castration-resistant prostate cancer (mCRPC) . Records were reviewed by humans and two AAMs. AAM-1 involved a pre-trained data-mining model specialized in biomedical literature, and AAM-2 was based on support vector machine algorithms. The AAMs assigned an accept/reject status, with reasons for exclusion. Automatic results were compared to those of humans. For mNSCLC, 5820 records were processed by humans and 440 (8%) records were accepted and the remaining items rejected. AAM-1 correctly accepted 6% of records and correctly excluded 79%. AAM-2 correctly accepted 6% of records and correctly excluded 82%. The review was completed by AAM-1 or AAM-2 in 52 hours, compared to 196 hours for humans. Work saved was estimated to be 76% and 79% by AAM-1 and AAM-2, respectively. For mCRPC, 2434 records were processed by humans and 26% of these were accepted and 74% rejected. AAM-1 correctly accepted 23% of records and rejected 62%. AAM-2 correctly accepted 20% of records and rejected 66%. The review was completed by AAM-1, AAM-2, and humans in 25, 25 and 85 hours, respectively. Work saved was estimated to be 61% and 68% by AAM-1 and AAM-2, respectively. AAMs can markedly reduce the time required for searching and triaging records during a systematic review. Methods similar to AAMs should be assessed in future research for how consistent their performances are in SLRs of economic, epidemiological and humanistic evidence.

2021 ◽  

Purpose: To assess the present landscape and future research directions, a bibliometric analysis was performed to identify the characteristics of the 100 most-cited articles (T100 articles) on CRPC research. Methods: A list of the T100 articles investigating CRPC was generated by searching the Web of Science (WoS) Core Collection database. Different characteristics of the T100 articles, including the countries/territories, journals, authors, and research areas, were analyzed. Results: The number of citations of T100 articles published between 1992 and 2017 ranged from 282 to 3594, with an average of 654.9 citations. According to the topic of the article, ''Mechanisms related to tumor progression or metastasis'' ranked first with 41 T100 articles, while immunotherapy ranked fourth with 7 T100 articles. The T100 articles originated from 31 countries, with more than half originating from the USA (n = 89). Professor Scher HI published the most T100 articles as the first author (4) and as the corresponding author (5), while Pro De Bono JS from the Institute of Cancer Research published 3 articles as the first author and 8 articles as the corresponding author. The journal Cancer Research published 20 T100 articles with a total of 8946 citations. The number of T100 articles(r = 0.485, P = 0.01) and the total number of citations(r = 0.626, P < 0.001) were all positively correlated with the IF of the journal. Conclusions: This analysis offers a historical perspective on the progress and attempts to reveal future trends in CRPC research using bibliometric analysis. This study's results suggest that immunotherapy and the study of androgen receptors as well as their signaling axes will possibly be hot topics and trends in CRPC research.


2020 ◽  
Vol 38 (6_suppl) ◽  
pp. 29-29
Author(s):  
Sebastien J. Hotte ◽  
Antonio Finelli ◽  
Kim N. Chi ◽  
Christina M. Canil ◽  
Neil Fleshner ◽  
...  

29 Background: The Canadian GU Research Consortium recently conducted a consensus development conference with 27 academic prostate cancer (PC) specialists leading to 31 consensus recommendations. We conducted a survey to compare community-based practice with the consensus recommendations on the management of metastatic castration sensitive prostate cancer (mCSPC), metastatic castration resistant prostate cancer (mCRPC) and non-metastatic castration resistant prostate cancer (nmCRPC). Methods: An 87-item online questionnaire was sent to 600 Canadian community urologists, medical oncologists, radiation oncologists, and general practitioner oncologists involved in the treatment of PC. Results: Seventy-two physicians responded to the questionnaire (12% response). A discordance of >25% was observed in 15 of 31 recommendations (48%). Among the areas of discordance were treatment approach for patients with nmCRPC and PSADT < 10 months who are negative for metastases on conventional imaging but metastatic on PET-based imaging. Of the academic physicians, 89% indicated treating with agents approved for nmCRPC compared to 50% of community physicians (p=0.0005). Important discrepancies were also observed across academic and community physicians for radiation to the prostate for low-volume mCSPC which was 74% vs 27%, (p<0.0001) respectively; criteria for stopping therapy in mCRPC in which 78% of academic physicians favored continuation of therapy in the event of PSA progression only, compared to 24% of community physicians. Sequencing of therapy after prior apalutamide for nmCRPC using subsequent docetaxel treatment was observed in 81% of academic physicians vs 35% of community physicians, (p<0.0001), and use of genetic testing was favored by 74% of academics vs 36% of community physicians, (p<0.0001) for newly diagnosed metastatic prostate cancer. Conclusions: The areas of discordance between a national sample of community-based PC physicians and academic consensus recommendations represent potential areas for education, practice tools and future research.


2020 ◽  
Vol 15 (2) ◽  
Author(s):  
Sebastien J. Hotte ◽  
Antonio Finelli ◽  
Kim N. Chi ◽  
Christina Canil ◽  
Neil Fleshner ◽  
...  

Introduction The Canadian Genito-Urinary Research Consortium conducted a consensus development conference leading to 31 recommendations. We conducted a survey comparing community-based practice with the consensus recommendations on the management of metastatic castration sensitive prostate cancer (mCSPC), metastatic castration resistant prostate cancer (mCRPC) and non-metastatic castration resistant prostate cancer (nmCRPC). Methods An 87-item online questionnaire was sent to 600 community urologists and oncologists involved in the treatment of prostate cancer.  Results Based on the 72 responses received, a discordance of ­>25% was observed in 15 recommendations (48%). Of the academic physicians, 89% indicated treating with agents approved for non-metastatic castration resistant prostate cancer compared to 50% of community physicians (p=0.0005). Discrepancies were also observed for radiation to the prostate for low-volume metastatic castration sensitive prostate cancer which was 74% (academic) vs 27% (community), (p<0.0001); criteria for stopping therapy in metastatic castration sensitive prostate cancer in which 78% of academic physicians favored continuation of therapy in the event of PSA progression only, compared to 24% of community physicians. Sequencing of therapy after prior apalutamide for non-metastatic castration resistant prostate cancer using subsequent docetaxel treatment was observed in 81% of academic physicians vs 35% of community physicians, (p<0.0001), and use of genetic testing was favored by 74% of academics vs 36% of community physicians, (p<0.0001) for newly diagnosed metastatic prostate cancer.      Conclusions The areas of discordance between a national sample of community-based physicians and academic consensus recommendations represent potential areas for education, practice tools and future research.  


2014 ◽  
Vol 32 (4_suppl) ◽  
pp. 185-185 ◽  
Author(s):  
Nicholas J. Vogelzang ◽  
Jeffrey L. Vacirca ◽  
Philip W. Kantoff ◽  
Mark C. Scholz ◽  
Shaker R. Dakhil ◽  
...  

185 Background: Sip-T is an autologous cellular immunotherapy indicated for asymptomatic or minimally symptomatic metastatic castration-resistant prostate cancer (mCRPC). PROCEED is an ongoing phase 4 registry, enrolling pts receiving commercial sip-T. There are no exclusions based on prior prostate cancer treatments, so sip-T product parameters can be evaluated in pts receiving prior hormonal agents such as ABI or ENZ. Methods: Pts treated with sip-T within ≤6 mo were eligible to provide informed consent. Prior anticancer interventions were recorded at baseline. Sip-T product parameters were assessed. Results: Of 1,376 pts (as of May 2013), 108 (7.8%) received prior ABI and 58 (4.2%) received ENZ. Patients with ABI or ENZ generally had more advanced disease vs. pts without prior ABI or ENZ (WPT) and a greater proportion of the ABI and ENZ pts received prior docetaxel. There were slight differences in some sip-T product parameters in the ABI and ENZ groups, but neither manufacture nor receipt of sip-T were inhibited, as indicated by the percentage of pts with 3 successful infusions. All groups showed evidence of immune prime boost (Table). Conclusions: A subset of patients in PROCEED have received ABI or ENZ prior to sip-T. We observed no impact of prior ABI/ENZ on sip-t manufacturing or delivery, despite these men having more advanced disease. These results are consistent with prior analyses demonstrating a positive sip-T treatment immune effect in patients who have received prior therapies, including chemotherapy. Further immune monitoring and outcome studies are currently ongoing in the pre-docetaxel mCRPC setting where these agents will likely be used clinically. Clinical trial information: NCT01306890. [Table: see text]


2016 ◽  
Vol 34 (2_suppl) ◽  
pp. 334-334
Author(s):  
Ajay Behl ◽  
Lorie Ellis ◽  
Dominic Pilon ◽  
Yongling Xiao ◽  
Patrick Lefebvre ◽  
...  

334 Background:Abiraterone acetate (ABI) and enzalutamide (ENZ) are novel oral therapies offering survival benefit to metastatic castration-resistant prostate cancer (mCRPC) patients. The efficacy of cancer treatments rely on patient consistency and adherence to recommended dosage regimens. Factors such as drug-drug interactions and intolerance or toxicities can result in patients or their providers reducing the drug dosage. This study aims to describe treatment discontinuation patterns observed for ABI and ENZ. Methods: The Truven Health MarketScan Research Databases were used to conduct a retrospective analysis of mCRPC patients initiated on ABI or ENZ (index date) between 10/01/2012 to 12/31/2014 with ≥6 months of continuous eligibility prior to index date and a PC diagnosis during the period of continuous eligibility. Patients were observed until loss to follow-up, or end of data availability. Kaplan-Meier (KM) survival curves were used to compare the rates of having a refill gap (i.e., ≥14 days, ≥30 days, or ≥60 days) in patients initiated on ABI or ENZ. Results: The table below summarizes the KM probabilities of having a refill gap among patients initiated on ABI or ENZ. Conclusions: Significantly higher rates of refill gaps ≥30 days and ≥ 60 days were observed for patients initiating ENZ as compared to ABI after 6 months of observation in this initial analysis. Additional research is needed to understand the reasons for gaps in treatment in patients treated with mCRPC therapies. [Table: see text]


Sign in / Sign up

Export Citation Format

Share Document