再発リンパ腫の治療における薬剤の組み合わせの個別化にQPOPが有効であることを確認(QPOP effective in personalising drug combinations to treat relapsed lymphoma)


標準治療にもかかわらず、がんが増殖する患者さんに、AIによる誘導治療が有効であることが判明 AI-guided treatments benefitted patients whose cancers grow despite standard therapy

2022-12-16 シンガポール国立大学(NUS)

 シンガポールの専門家が発表した新しい研究によると、患者固有の薬剤の組み合わせを特定する人工知能(AI)プラットフォームが、リンパ腫が再発した患者の助けとなることが示唆されています。10月19日に学術誌「Science Translational Medicine」に掲載されたこの論文は、リンパ腫患者に対する個別化された薬剤組み合わせ予測の実現可能性を示した初めての研究で、シンガポール国立大学(NUS)で開発されたQPOP(二次表現型最適化プラットフォーム)という新しい手法を利用しています。


再発・難治性リンパ腫における薬物併用療法を導くためのex vivoプラットフォーム An ex vivo platform to guide drug combination treatment in relapsed/refractory lymphoma

Jasmine Goh ,Sanjay De Mel,Michal M. Hoppe ,Masturah Bte Mohd Abdul Rashid,Xi Yun Zhang,Patrick Jaynes,Esther Ka Yan Ng,Nur’Atiqa Diana Binti Rahmat,Jayalakshmi,Clementine Xin Liu ,Limei Poon,Esther Chan,Joanne Lee,Yen Lin Chee,Liang Piu Koh,Lip Kun Tan,Teck Guan Soh,Yi Ching Yuen,Hoi-Yin Loi,Siok-Bian Ng,Xueying Goh,Donovan Eu,Stanley Loh,Sheldon Ng,Daryl Tan,Daryl Ming Zhe Cheah,Wan Lu Pang,Dachuan Huang,Shin Yeu Ong,Chandramouli Nagarajan,Jason Yongsheng Chan,Jeslin Chian Hung Ha,Lay Poh Khoo,Nagavalli Somasundaram,Tiffany Tang,Choon Kiat Ong,Wee-Joo Chng,Soon Thye Lim ,Edward K. Chow, Anand D. Jeyasekharan
Science Translational Medicine  Published:19 Oct 2022
DOI: 10.1126/scitranslmed.abn7824

Predicting possibilities with PDCs

Patients with non-Hodgkin’s lymphomas (NHLs) often relapse after frontline treatment, and interpatient heterogeneity make personalized combination treatment difficult. Goh et al. have developed a hybrid experimental-analytic method that they call quadratic phenotypic optimization platform, or QPOP, to identify personalized drug combination therapies using ex vivo patient samples to improve patient outcomes. In a prospective cohort, physicians were able to alter treatment according to drug combinations identified using QPOP after 6 days to achieve complete responses in 5 of 17 patients with NHL. This is a promising step for providing new hope for patients who have relapsed NHL and provides a foundation for further clinical trials.


Although combination therapy is the standard of care for relapsed/refractory non-Hodgkin’s lymphoma (RR-NHL), combination treatment chosen for an individual patient is empirical, and response rates remain poor in individuals with chemotherapy-resistant disease. Here, we evaluate an experimental-analytic method, quadratic phenotypic optimization platform (QPOP), for prediction of patient-specific drug combination efficacy from a limited quantity of biopsied tumor samples. In this prospective study, we enrolled 71 patients with RR-NHL (39 B cell NHL and 32 NK/T cell NHL) with a median of two prior lines of treatment, at two academic hospitals in Singapore from November 2017 to August 2021. Fresh biopsies underwent ex vivo testing using a panel of 12 drugs with known efficacy against NHL to identify effective single and combination treatments. Individualized QPOP reports were generated for 67 of 75 patient samples, with a median turnaround time of 6 days from sample collection to report generation. Doublet drug combinations containing copanlisib or romidepsin were most effective against B cell NHL and NK/T cell NHL samples, respectively. Off-label QPOP-guided therapy offered at physician discretion in the absence of standard options (n = 17) resulted in five complete responses. Among patients with more than two prior lines of therapy, the rates of progressive disease were lower with QPOP-guided treatments than with conventional chemotherapy. Overall, this study shows that the identification of patient-specific drug combinations through ex vivo analysis was achievable for RR-NHL in a clinically applicable time frame. These data provide the basis for a prospective clinical trial evaluating ex vivo–guided combination therapy in RR-NHL.