In The Lancet Oncology, Brian Rini and colleagues report the results of TIVO-3, a phase 3, multicentre, randomised, controlled, open-label trial comparing tivozanib, a selective VEGF receptor (VEGFR) tyrosine kinase inhibitor, with sorafenib, an earlier-generation tyrosine kinase inhibitor, as third-line or fourth-line therapy in patients with advanced, clear-cell, renal cell carcinoma. Tivozanib inhibits VEGFR1, VEGFR2, and VEGFR3, similar to axitinib, a drug previously compared with sorafenib as second-line therapy in mostly patients who progressed after VEGFR tyrosine kinase inhibitor or cytokine treatment. Progression-free survival in that study was 6·7 months with axitinib compared with 4·7 months with sorafenib (HR 0·66, 95% CI 0·54–0·81; p<0·0001). Patients had fewer lines of therapy than those in the study by Rini and colleagues, and no overall survival benefit was observed for axitinib.

We read with great interest the Article by Gianni Bisogno and colleagues, which showed that maintenance therapy improved survival in this group. As cited by the authors, the Soft Tissue Sarcoma Cooperative Group (Cooperative Weichteilsarkom Studiengruppe; CWS) has previously evaluated maintenance therapy in a non-randomised way in patients with metastatic disease. Additionally, the CWS-2002P study evaluated a similar maintenance therapy (vinblastine instead of vinorelbine) in an almost identically defined high-risk rhabdomyosarcoma patient group. We are also convinced that maintenance therapy could be an effective treatment for high-risk rhabdomyosarcoma and requires further evaluation. The results of the RMS 2005 study are, however, not entirely convincing. The primary endpoint of the study and primary goal for clinicians treating patients with high-risk rhabdomyosarcoma is to avoid relapse. In the study, this goal was not achieved; 5-year disease-free survival in the intention-to-treat population was 77·6% with maintenance therapy versus 69·8% without (p=0·061), and in the per-protocol analysis 77·8% vs 69·6% (p=0·053). This outcome is surprising as only one patient of 186 in the untreated therapy group, and three of 185 participants in the maintenance therapy group, were excluded in the per-protocol analysis. The 5-year overall survival was significantly higher in the therapy group (86·5%) compared with patients who stopped treatment (73·7%). As the authors stated, patient, tumour and relapse-related factors all influence survival after relapse and these factors were similarly distributed in both groups of the study. However, post-relapse therapy has a pivotal role for outcome and this factor could not be evaluated. The authors' hypothesis that “the combined regimen might have killed any residual tumour cells resistant to the drugs administered during the standard treatment” is not correct given the fact that relapse could not be significantly reduced by the maintenance therapy.

We read with interest the Article published by Gianni Bisogno and colleagues reporting the results of maintenance chemotherapy in children and adolescents with high-risk rhabdomyosarcoma. Despite the fact that the study was negative for its first endpoint, the authors suggest that “maintenance treatment with vinorelbine and low-dose oral cyclophosphamide for patients with high-risk rhabdomyosarcoma in complete remission after standard treatment” represents a new standard of care.

We welcome the comments of colleagues that allow us to further explain the addition of maintenance in the treatment of patients with high-risk rhabdomyosarcoma. Both letters state that the primary endpoint of our study, disease-free survival, was not satisfied. This result was thoroughly discussed in the paper, showing that disease-free survival was 77·6% with maintenance therapy and 69·8% when stopping treatment. Although this difference was not significant (p=0·061), this outcome, along with the significant improvement in overall survival, has led us to conclude that maintenance therapy should be included in treatment regimens.

Huiyan Luo and colleagues' multicentre, case-controlled, diagnostic study presented important data on the Gastrointestinal Artificial Intelligence Diagnostic System (GRAIDS) which, after training and validating over 1 million images via artificial intelligence, was shown to have a high diagnostic accuracy for detecting upper gastrointestinal cancers. The rapid pace of advancement in artificial intelligence has led to the development of autonomic technologies that assist doctors to make an accurate and quick decision.,  GRAIDS represents a tremendous opportunity to improve the accuracy of diagnosis and screening for upper gastrointestinal cancers. However, we are concerned about data interpretation and extrapolation in Luo and colleagues' study.

We appreciate Jialin Li and colleagues' comments on our recent study, which proposes use of the Gastrointestinal Artificial Intelligence Diagnostic System (GRAIDS) for endoscopic detection of upper gastrointestinal cancers. Li and colleagues questioned whether GRAIDS is equally effective across all stages of upper gastrointestinal cancers. Unfortunately, disease staging information was not collected for the past 10 years for the large-scale retrospective images obtained from the six participating centres. Nonetheless, T (tumour invasion) staging information was available in 386 patients and staging information from The American Joint Committee on Cancer was available in 428 patients in the prospective cohort. The accuracy of GRAIDS was consistently high for detecting upper gastrointestinal cancers across different stages (92·2%, 95·3%, 93·1%, and 95·4%, respectively, for T1, T2, T3, and T4 disease; and 93·6%, 92·7%, 95·3%, and 96·2%, respectively, for stage I, II, III, and IV disease).

In their report of quality-of-life measures in children and young adults with relapsed or refractory B-cell acute lymphoblastic leukaemia before and after a single infusion infusion of tisagenlecleucel in the ELIANA trial, Theodore W Laetsch and colleagues do not mention what proportion of patients received treatment with replacement immunoglobulin infusions as maintenance therapy after tisagenlecleucel. As an anti-CD19 chimeric antigen receptor therapy (CAR-T), treatment with tisagenlecleucel induces B-cell aplasia associated with hypogammaglobulinaemia as a direct pharmacodynamic consequence of its mechanism of action.

We are grateful to Siraj A Misbah amd Praveen Weeratunga for their comments on our study of patient-reported quality of life in children and young adults treated with tisagenlecleucel for B-cell acute lymphoblastic leukaemia. They correctly highlight that most patients in our ELIANA trial received immunoglobulin replacement and suggest that this might have contributed to the improvements in quality of life seen after treatment with tisagenlecleucel in our analysis of patients treated on this study.

We read with interest the study by Giovanni Ceresoli and colleagues testing Tumour Treating Fields (TTFields) for inoperable malignant pleural mesothelioma. TTFields is an anti-tumour technology delivering low-intensity electric fields that interfere with mitosis. This study on mesothelioma comes after other trials of TTFields in other advanced or recurrent tumours. Although the authors suggest that their results warrant further investigation, several points should be addressed. In particular, critical information is missing, such as history of asbestos exposure, precise staging of the disease, and other prognostic factors.

We read with great interest the paper by Giovanni Ceresoli and colleagues on the STELLAR study, a single-arm phase 2 study investigating Tumour Treating Fields (TTFields) in addition to standard first-line chemotherapy (platinum–pemetrexed) in malignant pleural mesothelioma. This new and innovative approach in malignant pleural mesothelioma has received US Food and Drug Administration approval under the Humanitarian Device Exemption pathway. However, in our opinion a few points require consideration before implementation of this strategy into daily practice.

We thank Marcella Barbarino and Cornedine Jannette de Gooijer and their colleagues for their interest in the STELLAR study. Barbarino and colleagues raise questions related to a possible selection bias in patient enrolment. Any single-arm phase 2 study has well known limitations. However, STELLAR was a relatively large, and well designed pilot study. Patients had incidental diagnosis of malignant pleural mesothelioma and were not part of a screening programme; the majority of them were enrolled in large-volume centres, located in regions with high occupational or environmental asbestos exposure. Patient baseline characteristics, including the percentage of patients with an Eastern Cooperative Oncology Group performance score of 0, were in line with other recent malignant pleural mesothelioma studies, such as the MAPS and LUME-Meso trials. Of note, 26% of patients in our study had a non-epithelioid tumour, a recognised negative prognostic factor captured by both CALGB and EORTC prognostic scores for malignant pleural mesothelioma.

We read with interest the article by Carrie and colleagues in The Lancet Oncology describing the long-term outcomes of the GETUG-AFU 16 trial. After a median follow-up of 112 months (IQR 102–123), metastasis-free survival was improved after 6 months of androgen deprivation therapy combined with salvage radiation therapy versus salvage radiotherapy alone in patients with biochemically recurrent prostate cancer without persisting prostate-specific antigen after radical prostatectomy (hazard ratio 0·73, 95% CI 0·54–0·98; p=0·0339). Patients were grouped by risk (of biological or metastatic recurrence or prostate cancer-related death, according to a combination of several risk factors (GETUG group criteria); however, metastasis-free survival did not differ between low-risk and high-risk subgroups.

We read with great interest the update to the GETUG-AFU 16 trial by Carrie and colleagues in The Lancet Oncology. We have a few comments regarding statements made and a request for additional information on the trial results.

We thank Pirus Ghadjar and Thomas Wiegel, and Samuel Birer and Daniel Spratt for their interesting comments on our GETUG-AFU 16 trial in The Lancet Oncology.

Corn PG, Heath EI, Zurita A, et al. Cabazitaxel plus carboplatin for the treatment of men with metastatic castration-resistant prostate cancers: a randomised, open-label, phase 1–2 trial. Lancet Oncol 2019; 20: 1432–43—In this Article, the National Institutes of Health and National Cancer Institute should have been included in the list of funders in the Summary. This correction has been made to the online version as of Jan 2, 2020.

In November, 2019, a Working Group of 13 scientists from eight countries met at the International Agency for Research on Cancer ( ) in Lyon, France, to finalise their evaluation of the carcinogenicity of five high- production-volume chemicals: glycidyl methacrylate, 1-butyl glycidyl ether, 1-bromo-3-chloropropane, 4-chlorobenzotrifluoride, and allyl chloride. These assessments will be published in volume 125 of the IARC Monographs.