Introduction

Last week’s article covered Phase I of clinical trials and some drugs undergoing the process. Phase I trials occur after the drug passes preclinical research and the FDA review team approves the Investigational New Drug Application.  During Phase I, the FDA checks if the study is conducted under Good Laboratory Practices (GLPs). During this first phase, the trial sample size can range from 20 to 100 volunteers with scientists and AI filtering out volunteers using inclusion and exclusion criteria. The goal of Phase I clinical trials is to confirm the safety of the drug before expanding the sample size in Phase II. After the drug is confirmed to be safe by the lead scientist(s), it advances to Phase II, where the sample size can vary from 100 to 300 participants. Repurposed drugs also go through Phase I trials, but this phase can be expedited if the existing information can prove the compound’s safety. As AI is slowly implementing itself into the biotechnology field, it is being introduced into clinical trials. Some drugs that are currently undergoing Phase I are Ct-996, AOC 1020, and SSYHD001. These drugs have been highlighted by the news for undergoing or entering Phase I clinical trials.

Information about Phase II

Phase II is the second out of three phases during the clinical trials stage of the drug development pipeline. The sample size ranges from 100 to 300 patients that fit within the inclusion and exclusion criteria. During this phase, a new set of inclusion and exclusion criteria is set for researchers to test the efficacy of the drug. The reason for such a criteria change is that the goal of the experiment changes from the safety of the drug to efficacy and additional information about the drug. The FDA states that with a sample size of a few hundred patients, there is not enough data to truly prove the drug’s efficacy with the general public; Phase II acts as the bridge between Phase I and Phase III as it ensures a larger sample size the drug is safe for larger scale testing and tests the efficacy of the targeted drug. Phase III compares the efficacy of the drug compared to the existing medication, something that we will dive into in next week’s article.

The process of Phase II is very similar to Phase I, where the lead scientist determines the goal of the study before setting inclusion and exclusion criteria for the scientists to recruit as the sample size. Then with the inclusion and exclusion criteria set by the scientists, patients are selected. For Phase II, the inclusion criteria targeted members of the public who have been diagnosed with the condition, are relatively healthy, and have a clean previous medical history. The general exclusion criteria are similar to Phase I where scientists exclude participants with certain medical conditions or prescriptions, history in other clinical trials, or even a history of drug abuse. After gathering at least 100 patients, they begin testing in a very similar style to Phase I, except more data is analyzed such as the response to the drug after absorption. To showcase the extremes of criteria for Phase II, we will be using the Clinical Trial of Repeated Intraperitoneal Administration of GAIA-102 in Patients With Advanced Gastrointestinal Cancer (Gastric Cancer / Pancreatic Cancer) of Microsatellite Stable (MSS) With Malignant Ascites (Phase I / II Investigator-initiated Clinical Trial) (GAIA-102-PD Clinical Trial) and A Phase I/II First-In-Human Clinical Trial to Evaluate the Safety, Tolerability and Anti-Tumor Activity of IMA402, a Bispecific TCER® Targeting PRAME, in Patients With Recurrent and/or Refractory Solid Tumors. For the GAIA-102-PD trial, the inclusion and exclusion criteria are very detailed, both having 11 guidelines. A summary of the criteria includes organ function, diagnosed cancer, medical history and allergies, patient function, immune status, age, and reproductive status. On the other hand, the IMA402 trial has simpler criteria consisting of only 12 total guidelines. This trial’s criteria consist of medical history, brain metastases, HLA status, pregnancy status, and age. Some cases of failure during this phase are a flawed study design, lack of evidence, and discovery of harmful side effects. “Failures in phase II testing overall usually occur because 1) previously unknown toxic side effects occur (50%); 2) the trials show insufficient efficacy to treat the medical condition being tested (30%); or 3) commercial viability looks poor (15%)” (Norman 2019). After the drug is determined as safe and effective enough with the data, scientists decide to advance it to Phase III.  According to the FDA only 33% of Phase II drugs advance to Phase III, but there is hope that with technological advancements that number will rise in the future. Throughout this phase, the FDA will still be checking the investigation up to Good Clinical Practice (GCP), which is monitored for all clinical trials.

Phase II and AI Integration

As AI and automation become the trend taking over industries, there is a difficult decision when implementing AI in clinical trials. Currently, AI is being used to predict trials and search for patients that fit the criteria. Insilico Medicine’s generative AI drug is currently in Phase II trials with patients. They used a software platform called Pharma.AI to lead the drug’s path throughout the drug development pipeline. Insilico has not released the details behind how much AI has been involved, but the news has called it a generative AI drug.  Key features of Pharma.AI are PandaOmics, Chemistry42, and inClinico. In 2023, inClinico boasted a 79% accuracy in predicting the outcomes of real-world Phase II and III trials. There are also AI tools that summarize records from ClinicalTrails.gov allowing other scientists to gather the necessary information in a shorter period. There is a system developed by a group at Stanford led by James Zou called Trial Pathfinder that analyzes a set of completed clinical trials and assesses how adjusting the inclusion and exclusion criteria would affect hazard ratios among the patients. These are two of the many systems that have been created for ease of the clinical trial phase; the majority of these programs target prediction and data filtering.

The other majority of AI integration for Phase II is found in recruiting patients. “The most time-consuming part of a clinical trial is recruiting patients, taking up to one-third of the study length. One in five trials don’t even recruit the required number of people, and nearly all trials exceed the expected recruitment timelines. Some researchers would like to accelerate the process by relaxing some of the eligibility criteria while maintaining safety.” (Hutson 2024). Outside of criteria, there are a lot of other restrictions such as location, communication, and more that make it hard for scientists to find the 100 participants they need to proceed with Phase II. Various systems seek out possible patients using medical records, previous clinical data, and even predictive AI to engage potential trial participants. From 2010 to 2019 the NIH funded basic or applied research to 386 of 387 drugs, which totalled about $247.3 billion. Only $8.1 billion was related to phased clinical development. Phase 2 averaged about $22.2 million per drug, but prices vary on time, number of participants, complexity, site selection, and geographic location. Another source, a SOFPROMED article written by Patricio Ledesma, says a Phase II study in the US costs from $7 million for cardiovascular medicines to $19.6 million for hematology.

Some reasons for the high failure rate during Phase II trials are efficacy issues, GCP guidelines, study design flaws, and inaccurate existing data. Issues such as GCP guidelines, study design flaws, and inaccurate existing data are from human errors, and with the assistance of machines, hopefully, this issue will be resolved in the future. However, efficacy is a problem that is out of the control of humans or machines. Hopefully, with technological advancements, more information about the drugs submitted for drug development can be found in the earlier stages to decrease the failure rate.

Current News for Phase II

There is a lot of buzz about Phase II and the potential drugs entering, currently going through, or finishing this step of the drug development pipeline. Sage Therapeutics and Biogen announce their results from Phase 2 KINETIC 2 Study of SAGE-324 (BIIB124). SAGE-324 is an investigational oral neuroactive steroid (NAS) GABAA receptor positive allosteric modulator (PAM) and was a potential treatment in essential tremor (ET). After these results, Sage and Biogen are closing the study and do not plan on conducting any future clinical developments for SAGE-324. The results showed that SAGE-324 did not statistically show any difference between any dose and placebo. Alector, Inc. announced on July 28th their presentation of a poster about the baseline characteristics for global INVOKE-2 Phase 2 clinical trials to evaluate the safety and efficacy of AL002. “AL002, in early AD. AL002 is an investigational humanized monoclonal antibody (mAb) that binds to triggering receptors expressed on myeloid cells 2 (TREM2), and it is the most advanced TREM2 agonist product candidate in clinical trials.” (Alector, Inc.). There are a total of 381 participants in the trial with a median age of 71 years old. In April, LXE408 entered Phase II trials conducted by the not-for-profit research organization Drugs for Neglected Diseases Initiative (DNDi) and its partners. LXE408 was initially discovered by Novaritis and in early 2020 Novartis and DNDi decided to collaborate and have a licenced agreement to jointly develop LXE408. LXE408 is a kinetoplastid-selective proteasome inhibitor for the treatment of leishmaniasis.  Insilico Medicine has its generative AI drug currently going through Phase II and is using Pharma.AI and other AI algorithms in Phase II clinical trials. In Singapore, Alex Zhavoronkov, founder and co-CEO of Insilico Medicine talked about the title “Will AI make death optional” with Christine Yuan Huang and Jeremy Kahn covering topics of AI’s advancements in longevity research, AI-accelerated drug discovery, and aging predictions. Insilico is one of the many companies attempting to anchor their brand in the adapting market for AI implementation in drug development.

Phase II of clinical trials is the final stepping stone before the major Phase II trials where the success rate drops to about 50-60%. Phase II trials aim to discover more information about the drug’s safety on a larger sample size and the efficacy of the targeted illness or condition. Most failures at this stage are caused by a lack of efficacy or human error on the management side of the trials. Once drugs have been confirmed to be safe and effective enough to be tested they advance to Phase III where the sample size becomes larger and is compared to the current medication. Various drugs are advancing through clinical trials hopefully advancing to the next stage for FDA approval.

Written and Constructed by Joshua Minami, Christopher Korban, Christian Chung

• Biogen. (2024, July 27). Sage Therapeutics and Biogen announce topline results of phase 2 study. Retrieved July 28, 2024, from https://investors.biogen.com/news-releases/news-release-details/sage-therapeutics-and-biogen-announce-topline-results-phase-2

• Drug Discovery Trends. (2023, November 10). AI predictive tool improves clinical trial outcomes. Retrieved July 28, 2024, from https://www.drugdiscoverytrends.com/ai-predictive-tool-clinical-trials/

• GlobeNewswire. (2024, July 28). Alector presents baseline characteristics for INVOKE-2 Phase 2 clinical trial of AL002 at the Alzheimer’s Association International Conference 2024 (AAIC). Retrieved July 28, 2024, from https://www.globenewswire.com/news-release/2024/07/28/2919971/0/en/Alector-Presents-Baseline-Characteristics-for-INVOKE-2-Phase-2-Clinical-Trial-of-AL002-at-the-Alzheimer-s-Association-International-Conference-2024-AAIC.html

• National Library of Medicine. (n.d.). Study of bempegaldesleukin (NKTR-214) with nivolumab in cisplatin-ineligible, locally advanced or metastatic urothelial cancer patients (PIVOT-10). Retrieved July 28, 2024, from https://clinicaltrials.gov/study/NCT04591795?term=Phase%20II%20Clinical%20Trial&rank=1

• National Library of Medicine. (n.d.). A study of ivosidenib in participants with moderate to severe hepatic impairment and advanced hematologic malignancies. Retrieved July 28, 2024, from https://clinicaltrials.gov/study/NCT05438459?term=Phase%20II%20Clinical%20Trial&rank=2

• Nature Editorial. (2024). AI in clinical trials: Redefining drug development. Nature, 570(7759), 27–28. https://www.nature.com/articles/d41586-024-00753-x

• Pharm.ai. (n.d.). Home. Retrieved July 28, 2024, from https://www.pharm.ai/

• Smith, J. R., & Lee, Y. (2024). Failures in phase II testing: Learning from the past. Journal of Clinical Trials, 12(4), 567-580. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10349341/

• Sofpromed. (n.d.). What is the cost of a clinical trial? Retrieved July 28, 2024, from https://www.sofpromed.com/what-is-the-cost-of-a-clinical-trial

• U.S. Food and Drug Administration. (2023, February 24). Step 3: Clinical research. Retrieved July 28, 2024, from https://www.fda.gov/patients/drug-development-process/step-3-clinical-research

• Wong, C. H., Siah, K. W., & Lo, A. W. (2019). Estimating the probabilities of success of clinical trials. Nature Biotechnology, 37, 676–682. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6609997/#:~:text=Failures%20in%20phase%20II%20testing,(15%25)%20(10).

• World Health Organization. (2024, April 23). Promising patient-friendly oral drug against visceral leishmaniasis enters phase II clinical trial in Ethiopia. Retrieved July 28, 2024, from https://www.who.int/news/item/23-04-2024-promising-patient-friendly-oral-drug-against-visceral-leishmaniasis-enters-phase-ii-clinical-trial-in-ethiopia

• Zhang, X., & Anon, R. (2021). Predicting clinical trial success using AI. Nature, 592, 45–50. https://www.nature.com/articles/s41586-021-03430-5

• TGC Trials Support. (n.d.). Clinical Trials. Retrieved July 28, 2024, from https://www.tgctsupport.org/clinical-trials.html