Scientific corner

Machine learning using clinical data at baseline predicts the medium-term efficacy of ustekinumab in patients with ulcerative colitis

Hiromu Morikubo, Ryuta Tojima, Tsubasa Maeda, Katsuyoshi Matsuoka, Minoru Matsuura, Jun Miyoshi, Satoshi Tamura & Tadakazu Hisamatsu Sci Rep 14, 4386 (2024).

Predicting the therapeutic response to biologics before administration is a key clinical challenge in ulcerative colitis (UC). We previously reported a model for predicting the efficacy of vedolizumab (VDZ) for UC using a machine-learning approach. Ustekinumab (UST) is now available for treating UC, but no model for predicting its efficacy has been developed. When applied to patients with UC treated with UST, our VDZ prediction model showed positive predictive value (PPV) of 56.3% and negative predictive value (NPV) of 62.5%. Given this limited predictive ability, we aimed to develop a UST-specific prediction model with clinical features at baseline including background factors, clinical and endoscopic activity, and blood test results, as we did for the VDZ prediction model. The top 10 features (Alb, monocytes, height, MCV, TP, Lichtiger index, white blood cell count, MCHC, partial Mayo score, and CRP) associated with steroid-free clinical remission at 6 months after starting UST were selected using random forest. The predictive ability of a model using these predictors was evaluated by fivefold cross-validation. Validation of the prediction model with an external cohort showed PPV of 68.8% and NPV of 71.4%. Our study suggested the importance of establishing a drug-specific prediction model.

Scientific corner

Tu1846 Higher TNFα and Mip-1β Expression in Pretreatment Colonic Mucosa Have Potential to Predict of Achieving Mucosal Healing by Granulocyte and Monocyte Adsorptive Apheresis Therapy in Ulcerative Colitis Patients

Chie Kurihara, Toshihide Ohmori, Hirotaka Furuhashi, Kenichi Inaba, Nao Sugihara, Yoshinori Hanawa, Gastroenterology 2019 156 (6) Suppl. S-1146

Background: Granulocyte and monocyte adsorptive apheresis (GMA) is non-pharmacological therapy which selective depletion of activated granulocytes and monocytes/macrophages from peripheral blood. GMA is effective and safe as induction therapy in ulcerative colitis (UC) of moderate to severe patients, and commonly used in Asia and north Europe. However, therapeutic mechanism of GMA, especially its effect on mucosal healing, has not been well characterized. Since moderate to severe patients with UC sometimes become fulminant, it is important to select an appropriate induction therapy. GMA treatment efficacy is reported about 60% patients respond, whereas, there is not useful parameter that predicts GMA therapeutic efficacy before treatment. In this study, we attempted to identify predictive factors of clinical response to GMA treatment in UC patients. Methods: Thirty-two active UC patients (Mayo score ^ 5 and Mayo endoscopic score ^ 2) and 10 non-IBD control subjects were enrolled in this study. All UC patients received 10 or 11 times of GMA, and colonoscopies were applied before the first GMA and after the last GMA. Control subjects underwent colonoscopies for screening of colon cancer. Assessment of disease activity and colonic mucosal healing were determined based on Mayo score. Inflammation-related molecules mRNA expressions were determined by quantitative RT-PCR using biopsy specimen of colonic mucosa. ROC curves analysis was used to assess sensitivity and specificity in prediction of GMA therapeutic efficacy. Results: GMA treatment efficacy is 11 patients (34.4%) achieved clinical remission, 17 patients (53.1%) were response and 4 patients (12.5%) were non-response. Mucosal healing was observed in 18 patients (56.3%) and it was not observed in 14 patients (43.7%). Baseline characteristics were not significantly different according to GMA efficacy. Before the first GMA session, the clinical remission group showed significantly higher expressions of TNFα, MAdCAM-1 and MIP-1β mRNA than those of the non-response group (P <0.05). Patients in the response group who highly expressed these mRNA achieved mucosal healing. In the mucosal healing group, the mRNA levels of TNFα, IL-1β, IL-8, MIP-1β, TGFβ and IL-10 were significantly higher than those in the non-mucosal healing group (P <0.05). TNFα and MIP-1β had 0.83 and 0.79 of area under the curve with 83.3% and 66.7% sensitivity, 71.4% and 100% specificity, 78.9% and 100%positive predictive value, and 76.9% and 70.0% negative predictive value, respectively, for prediction of mucosal healing. Conclusion: Mucosal expression of TNFα and MIP-1β mRNA before treatment in the remission group and the mucosal healing group was significantly higher than non-responder to GMA treatment. We propose that measuring of these molecules’ expression is useful to expect GMA therapeutic efficacy in patients with UC.

Scientific corner

Predictors of clinical response in patients with ulcerative colitis treated with granulocyte-monocyte apheresis: analysis of the apheresis registry data.

Pietro Manuel Ferraro 1Valeria D’OvidioGiampaolo BresciMarco AstegianoBeatrice PrincipiRenata D’Inca’Roberto TestaDaniela ValpianiLuisa GuidiAlessandro ArmuzziFrancesco CostaAnnalisa AratariMaurizio VecchiRoberto De FranchisVincenza Di LeoChiara RicciGabriele RieglerElisabetta ColomboGiuseppe RepaciPierenrico LecisMichele SillaStefano Passalacqua, G Ital Nefrol. Jan-Feb 2012;29 Suppl 54:S144-6.

We analyzed predictors of clinical response after a cycle of granulocytemonocyte apheresis in 173 patients with ulcerative colitis. Hemoglobin levels independently predicted good clinical outcome.

Scientific corner

A retrospective search for predictors of clinical response to selective granulocyte and monocyte apheresis in patients with ulcerative colitis

Yasuo Suzuki 1Naoki YoshimuraKatsuyuki FukudaKoji ShiraiYasushi SaitoAbbi R Saniabadi

Dig Dis Sci.  2006 Nov;51(11):2031-8. doi: 10.1007/s10620-006-9199-9. Epub 2006 Sep 27.

Recently, selective granulocytapheresis (Adacolumn) has appeared as a new treatment for patients with inflammatory bowel disease. This study sought to determine predictors of response to this new nonpharmacologic mode of therapy by retrospectively evaluating 28 patients who received granulocytapheresis after experiencing active ulcerative colitis (UC). Between April 2000 and March 2004, 28 consecutive patients received granulocytapheresis for active UC with the Adacolumn, which is filled with cellulose acetate beads as the column leukocytapheresis carriers; the carriers adsorb granulocytes, monocytes/macrophages, and a small fraction of lymphocytes (FcgammaR and complement receptors bearing leukocytes). Each patient could receive up to 10 Adacolumn sessions, at 2 sessions per week. In 2004, clinical response was retrospectively evaluated. Seven days after the last Adacolumn session, 20 of 28 patients had remission (colitis activity index [CAI] < or =4) including all 8 patients who had their first UC episode. The mean duration of UC in the 8 first episode cases was 3.4 months compared with 40.2 months for all 28 patients and 65.4 months for the 8 nonresponders. The response to Adacolumn was independent of basal CAI. The 8 nonresponders were given conventional medication (CM) or cyclosporine (CsA) if the former failed. Two responded to CM, 3 to CsA, and 3 underwent colectomy. First UC episode and short disease duration appear good predictors of response to granulocytapheresis. Selective granulocytapheresis might be an effective first-line treatment.

Contact UsFor more information

Contact Us