Last week’s round-up, 02-06 September 2019

By Tessa Fiorini-Cohen and Marthese Mifsud


Has AI Discovered a Drug Now? Guess.

06 Sep 2019

The media is abuzz with reports that an artificial intelligence model has designed a drug in 21 days, however an opinion piece in Science Translational Medicine takes a more nuanced view.

The software model is acronymed GENTRL (generative tensorial reinforcement learning), and this generates potential new structures for medicinal compounds using a multilayer optimisation approach. This allows the software to propose structures that are likely to hit a targeted receptor, whilst not hitting other related receptors and not resembling structures that are already patented. A number of filtering rounds then clears out various identified compounds based on molecular weight, instability, reactivity, and more, to produce a restricted set of potential compounds for eventual experimental validation.

This opinion piece reviews the technology in further detail and concludes that it is a very interesting step forward, but that virtual screening still needs to get much better before it can be used to successfully discover drug candidates:


GAO Pushes FDA to Improve First Cycle Generic Approval Rate

05 Sep 2019

The United States’ Government Accountability Office (GAO) has called on the FDA to improve its first cycle generic approval rate.

While the first cycle approval rate for new drugs is around 90%, that for generics is only 12%. It was noted however that this approval rate is a marked increase from 2012, when it stood at less than one percent.

The GAO specifically pointed out the need for improvements in clarity and consistency of reviewer comments. Based on industry feedback, it has also asked the FDA to assess the extent to which the timing of brand-name drug companies’ labelling changes affects the first cycle generic approval rate, and to take steps to limit this effect as appropriate:


No-Deal Brexit: How Medicines and Devices Will be Regulated in the UK

04 Sep 2019

With Brexit’s October deadline drawing near, the MHRA has reiterated guidance on what will happen to drug and medical device regulations in the case of a hard Brexit.

The guidance lays out what will happen to centrally authorised products, generic applications, parallel imports, medical device regulation, and clinical trials regulation within the UK. Changes in safety reporting, paediatric investigation plans, and designation of orphan medicines are also covered. Besides explaining how the agency itself is preparing for Brexit, the guidance details the work required from marketing authorisation holders to keep their medicines and medical devices on the UK market. Similarly, the work required from wholesalers and sponsors of clinical trials is also covered:


The latest issue of the EMA’s SME Office Newsletter has been released.

03 Sep 2019

This newsletter outlines the new scientific guidelines that have, or will soon, come into force. Included amongst these are the guideline on sterilisation of active substances, excipients, medicinal products and primary containers, as well as the guideline on subgroups analyses in confirmatory clinical trials included in marketing authorisation applications.

The newsletter also highlights new Q&As that have been published and guidelines that have been released for public consultation. Included amongst these are Q&A on bioequivalence studies with crushed tablets for generic applications, and procedural Q&As on paediatric investigation plans, legal establishment of marketing authorisation applicants in the EU, risk management plans, classification of post-authorisation changes, post-authorisation measures and extensions of marketing authorisations. Guidelines open for consultation include ones on quality requirements for medical devices, bioanalytical method validation, and general considerations for clinical trials:


FDA Clarifies Impurity Risks From ARB Recalls

02 Sep 2019

The FDA has reduced its initial risk estimates associated with last year’s discovery of nitrosamine impurities in medicines containing sartans.

The Agency’s initial worst case estimate was one additional case of cancer in 8,000 people taking the medicine for four years. However this calculation was based on all affected patients taking the highest possible medication dose, assuming the consistent presence of N-Nitrosodizzmethylamine (NDMA). In reality, most patients will not always have received a contaminated product and, when they did, they would have been exposed to much smaller amounts of the impurity.

The FDA is continuing to work with global regulatory agencies to deepen joint understanding of the issue, and will be using these findings to inform its evaluation of medicines manufactured similarly:


Image: Artificial Intelligence & AI & Machine Learning, CC BY 2.0


Trackbacks and pingbacks

No trackback or pingback available for this article.