Oil fell as the dollar strengthened and traders doubted US President Donald Trump’s plan to pressure Moscow would disrupt Russian exports.
West Texas Intermediate slid 0.7% to settle above $66, extending Monday’s losses. Trump told reporters the US will impose a 19% tariff on goods from Indonesia after teasing the deal earlier in the day. The dollar strengthened, making commodities priced in the currency less attractive.
As a child, Demis Hassabis, the winner of the 15th UKtech50, was already somewhat unique. A chess prodigy from the age of four, he taught himself programming on a ZX Spectrum 48K computer by the time he was eight. He began designing and writing video games, gaining his first taste for artificial intelligence (AI), as a teenager, which would later lead to receiving a Nobel Prize and a knighthood in the same year for his work.
Hassabis wears many hats: chess player, video game designer, neuroscientist, researcher and – most notable – AI entrepreneur. Perhaps best known for co-founding DeepMind in 2010, together with friends Shane Legg and Mustafa Suleyman, an AI startup which was acquired by Google four years later for around £400m, his knowledge of both science and technology has served him well.
The two topics go hand in hand. After graduating from Cambridge with a degree in computer science, Hassabis went on to do a PhD in cognitive neuroscience at University College London, studying how memory and imagination are linked to the brain. That thought was what led to the founding of DeepMind.
In a 2022 TED interview, Hassabis said that the link between the dopamine response in a human brain is replicated in AI, in that finding the right action that will best lead towards the overall goal.
“One of the things that has been so interesting about the convergence of some neuroscience and AI over the past 20 or 30 years is our understanding of that reward mechanism – the dopamine mechanism that we talk about in the brain,” he said.
“The popular explanation of it is that dopamine response to reward in the external world, but in fact, it responds to expectations about reward, right? You’re imagining that you’re gonna get $5 and then you get $10, and so there’s a dopamine surge because you exceeded expectations, and vice versa. And that turned out to be relevant in the world of AI as well.”
It was while at UCL that Hassabis met Suleyman and Legg, and the three came up with the idea that neuroscience, AI and machine learning together could create powerful algorithms, thus creating DeepMind.
One of the company’s first early achievements, before it was bought by Google, was using deep reinforcement learning to create an unbeatable AI model that could win Atari games such as Space Invaders and Pong. DeepMind then moved onto more complex games, gaining the attention of tech giant Google.
Following the Google acquisition, Hassabis stayed on as DeepMind’s CEO, and in 2016, the company launched its AI system AlphaGo, which famously defeated world champion Lee Sedol in the complex, ancient board game of Go.
This showed, for the first time, the true potential of general AI to discover solutions humans may not have even considered, not just in the field of video games, but in areas including medicine and science.
AlphaGo itself, as well as the concept of artificial general intelligence (AGI), led to the development of AlphaZero, a game-playing system which taught itself to master chess as well as other games, without any human guidance. However, the journey of DeepMind’s AI systems has since moved on.
As well as advancing research on AI safety and the development of a partnership with London’s Moorfields Eye Hospital for the use of artificial intelligence to identify and treat degenerative eye conditions, DeepMind developed AlphaFold in 2018, which is a system to solve the protein folding problem.
The system accurately predicted the 3D shape a protein will fold up to when it’s in the body, a concept which was first articulated by another Nobel Prize winner, Christian Anfinsen, in 1972. DeepMind used 150,000 proteins whose structure had already been identified to train AlphaFold to predict their shape. This then led to the development of Alphafold2, for which Hassabis won a Nobel Prize.
It was 2020 when Hassabis, together with DeepMind colleague, American chemist and computer scientist John Jumper, created the second version of the AI model for structure prediction. The model can accurately predict the 3D model of protein structures by taking the protein’s amino acid sequence, not only on single protein chains. There are currently more than 200 million predicted protein structures available in the AlphaFold database, a far cry from the 150,000 that were known before Hassabis created the AI model.
The AlphaFold database is currently being used by several sectors all over the world, including pharmaceutical research, healthcare and environmental technology. In Singapore, researchers are using it to come up with ways to diagnose and treat Parkinson’s disease quicker, while the US is using it to combat antibiotic resistance. In Norway, researchers have used it to figure out how to increase honeybees’ chance of survival through looking into vitellogenin, a protein fundamental to bees’ immune system. Meanwhile, researchers at the University of Portsmouth in the UK have screened 100 candidate enzymes to engineer faster and cheaper plastic-recycling enzymes.
So, what’s next for Hassabis and AI? Despite progress, he believes there is still a while to go before many of the capabilities of AI will have evolved to live up to its current hype.
In March 2025, speaking at an event to mark the availability of audio generation model Chirp 3 on the Google Vertex AI platform, Hassabis said that AI will have evolved to AGI, whereby the AI system exhibits “the cognitive capabilities” of humans, within that time. “That’ll be a moment when we have finally arrived with a kind of general intelligence, which is the original aim of the whole field of AI,” he added.
While AI models have evolved drastically in the past few years, there are still challenges when it comes to combining them with planning algorithms. “If your AI model has a 1% error rate and you plan over 5,000 steps, that 1% compounds like compound interest,” Hassabis said at the event.
By the time 5,000 steps have been worked through, the compounded error, according to Hassabis, means the possibility of the answer being correct is random. “For a games model, you have the rules of Chess or Go,” he said, which aids the planning algorithm in making the correct decision. “In the real world, you don’t have perfect information. There’s hidden information that we don’t know about, so we need AI models that are able to understand the world around us.”
For Hassabis, one of the interesting developments expected to appear over the next few years is the deployment of multiple AI agents that work together to solve a problem, and AI agents themselves can be used in a general AI system to solve problems.
Only a few weeks ago, DeepMind launched its latest venture, the AlphaGenome tool, which accurately predicts how single variants or mutations in DNA sequences affect the biological processes regulating genes.
However, this too has limitations when it comes to accurately capturing the “distant regulatory elements, like those over 100,000 DNA letters away, [which is] still an ongoing challenge”, according to DeepMind. “Another priority for future work is further increasing the model’s ability to capture cell- and tissue-specific patterns.”
The North American summer hard-court swing will kick off with a sparking field at the WTA 500 Mubadala Citi DC Open in Washington, D.C.
The main-draw entry list for this year’s event was released the week of June 23rd. The main-draw field (including wild cards) currently features three Top 10 players and five Grand Slam champions. Main-draw play in Rock Creek Park kicks off on Monday, July 21.
World No. 4 Jessica Pegula tops the field in her nation’s capital. Pegula won her first career WTA singles title at Washington in 2019 ranked No. 79, which kickstarted her rise up the PIF WTA Rankings.
Pegula has been a regular Top 5 player over the last few years and has already won three WTA singles titles in 2025, including last month on the grass of Bad Homburg.
As of the July 14 rankings, the other Top 10 players on the main-draw entry list are another American, Wimbledon finalist Amanda Anisimova, and Zheng Qinwen of China, a late addition to the field.
Anisimova made her Top 10 debut on Monday following a run to her first Grand Slam final at Wimbledon.
Unfortunately, another Top 10 player, defending champion Paula Badosa, withdrew from action for the next few weeks due to a back injury. Last year in Washington, former World No. 2 Badosa won her first title since 2022.
The Grand Slam champions named on the entry list are three titlists from this decade: 2021 US Open champion Emma Raducanu, 2022 Wimbledon champion Elena Rybakina and 2020 Australian Open champion Sofia Kenin.
The tournament also announced that a fourth Grand Slam champion, Naomi Osaka, would be at the event as well. Osaka has won four Grand Slam titles: 2018 and 2020 US Open, and 2019 and 2021 Australian Open.
Also, seven-time Grand Slam champion Venus Williams will join the field as a main-draw wild card. The former World No. 1 will return to Hologic WTA Tour action for the first time in 16 months.
Other Top 25 players on the entry list include Emma Navarro, Ekaterina Alexandrova, Clara Tauson and Magdalena Frech.
The ATP will also be hosting its own 500-level tournament side by side with the women’s event. This is the only combined WTA-ATP 500-level tennis tournament on the tennis calendar all year.
The full main-draw player field can be found at the Mubadala Citi DC Open website here!
R/R AML | Image credit:
© Oleksandr – stock.adobe.com
Final results from a single-institution phase 2 study (NCT04517851) demonstrated that treatment with single-agent elotuzumab (Empliciti) was safe and efficacious in patients with myelofibrosis who were not candidates for JAK inhibitors, supporting further investigation of the agent.
Findings presented at the 2025 EHA Congress showed that in treated patients (n = 15), the vast majority of adverse effects (AEs) possibly related to elotuzumab were grade 1 or 2. Grade 1 AEs possibly related to treatment included acute pain during infusion (n = 1), bone pain (n = 1), diarrhea (n = 4), dizziness (n = 2), flushing (n = 1), headache (n = 2), paresthesia (n = 1), pruritus (n = 1), voice alteration/hoarseness (n = 1), increased alanine aminotransferase (ALT) levels (n = 1), and hyper glycemia (n = 1).
Grade 2 AEs possibly related to treatment included fever (n = 1), chills as an infusion-related reaction (n = 2), fever as an infusion-related reaction (n = 1), pharyngitis (n = 1), and elevated ALT levels (n = 1). Grade 3 AEs possibly related to treatment comprised diarrhea (n = 1) and elevated aspartate aminotransferase levels (n = 1).
Regarding efficacy, patients achieved an overall response rate (ORR) of 33%; responses comprised a clinical improvement in anemia (n = 2), major platelet response (n = 1), clinical improvement in symptoms (n = 1), and both clinical improvement in symptoms and major platelet response (n = 1).
“Elotuzumab is active in myelofibrosis and has an excellent safety profile,” lead study author Prithviraj Bose, MD, and colleagues wrote in a poster presentation of the data. “Combination studies with JAK inhibitors may be warranted.”
Bose is a professor in the Department of Leukemia of the Division of Cancer Medicine at The University of Texas MD Anderson Cancer Center in Houston.
Investigators enrolled adult patients with primary myelofibrosis or post–polycythemia vera/ essential thrombocytosis myelofibrosis who were not candidates for treatment with a JAK inhibitor, based on the opinion of the treating physician. Notably, prior treatment with a JAK inhibitor was permitted. Patients were required to have disease harboring JAK2 V617F mutations, and intermediate- or higher-risk disease was needed.2 An ECOG performance status of 0 to 2 or a Karnofsky performance status of at least 60% was also required for enrollment.
All enrolled patients received elotuzumab intravenously at 10 mg/kg once per week for 8 weeks, then 20 mg/kg once every 4 weeks.1 Treatment continued for a maximum of 36 cycles, or until disease progression or unacceptable toxicity.
ORR served as the trial’s primary end point.2 Secondary end points included safety, duration of response (DOR), time to next treatment, complete response rate, partial response rate, rate of clinical improvement, platelet response rate, and changes in bone marrow fibrosis grade.
At the time of the final analysis, 1 patient remained on study treatment.1 Among the 14 patients who discontinued treatment, reasons for discontinuation included no response (n = 8), completing 36 cycles of treatment (n = 4), leukemia transformation (n = 1), and death from gastrointestinal bleeding unrelated to elotuzumab (n = 1). Four patients from the population died, including 1 during the study.
Findings also showed that the median time to response was 2.8 months (range, 0.1-3.7), and the median DOR was 11.1 months (range, 3-32.6). The median overall survival was not reached (range, 5.6-38).
The median treatment duration was 6.6 months (range, 0.9-35.4), and patients received a median of 7 cycles of treatment (range, 1-36). The median follow-up was 26.9 months (range, 10.3-38) in living patients.
Notably, 1 patient experienced a transformation from marrow fibrosis grade 3 to grade 2 after 2.8 months of treatment, and this was sustained through 36 cycles of treatment. Another patient experienced a hemoglobin level increase of more at least 2 g/dL; however, this patient was off the study prior to week 12.
Jerry McGuire, MD
Credit: University of California, Riverside
At the 2025 Southern California Psychiatry (So Cal Psych) Conference from July 11 – 24 at Huntington Beach, California, Jerry McGuire, MD, director of the residency training and chair of psychiatry at the College Medical Center in Long Beach, California, presented on how to detect bipolar depression from major depressive disorder (MDD). It comes down to 1 thing: family history.
McGuire is also the founder and president of the Stuttering Treatment and Research Society (STARS), a non-profit organization that has raised over $10 million to help people who stutter and raise awareness for this neurological condition. In this interview, McGuire shared how to differentiate bipolar depression from MDD, causes and treatments for agitation, and how to treat anxiety or depression caused by stuttering.
HCPLive: Today at the conference, you’re talking about the difference between bipolar depression and MDD. Can you tell our audience how they tell the difference?
McGuire: The key is cross-sectional. When you see someone and they’re depressed, you can’t tell necessarily. I wish in psychiatry we had blood tests or biomarkers, but [it] wouldn’t be as much fun.
The best way we differentiate is by history, and sometimes not just from the patient, but from the family members or others around them. We look longitudinally back; if there’s ever been a history of mania, then this depressive episode that we’re looking at now would be a bipolar I depression. If there’s history of only hypomanic episodes, and now we’re looking at a depressive episode, then that would be bipolar 2 depression.
There are certainly other clues in our diagnostic schema. One, if a patient has had a… misadventure on an antidepressant, if they’ve been prescribed an antidepressant and they got worse, they got more agitated…that may be a clue that what you’re dealing with is a bipolar depression. If there’s a family history, unfortunately, of people passing from suicide, that may be a clue as well, and even sometimes coexistent substance use that patients with bipolar may, at times, self-medicate. It’s important to look into these other co-existing factors that may give us clues that this is a bipolar depression, as opposed to a major depressive episode of major depressive disorder.
HCPLive: Once you’ve made that determination, what kind of treatments are you going to consider?
McGuire: We tailor our therapies based upon the diagnosis. One thing I’ve always tried to teach my residents, if your known treatments aren’t working, is a diagnosis correct? Sometimes [after] 3 or 4 antidepressants… patients aren’t responding, or they’re getting worse. Is it really major depressive disorder? That’s 1 clue.
Keep in mind that not many agents are approved for bipolar depression. Most of our agents that we talk about as mood stabilizers may be effective in preventing a manic episode or treating mania or mixed states. Our options are more limited when it comes to bipolar depression, and we do have some agents that are FDA-approved specifically for bipolar depression.
HCPLive: You’re also talking about agitation and dementia, which has been a big topic lately, right?
McGuire: With agitation and dementia, our hands have been really tied as clinicians, because unfortunately, for years, patients [were] getting older, and we had nothing FDA approved to treat them. If anything, agents we were using were given a box warning not to use about the risk potentially of increased mortality, cerebrovascular adverse events with antipsychotic agents in patients with agitation associated with dementia.
Fortunately, we have one agent that is on label now, and that’s been FDA approved for the treatment of patients with agitation associated with dementia due to Alzheimer’s. It’s nice to know that we have that option. Fortunately, it appears that we may have other options, BFD approved in the fairly recent future.
HCPLive: Could there be something that’s leading to the agitation? Has a patient, say, had a change in environment?
McGuire: When we change the environment [of patients with dementia], they may become more restless, more agitated. Look for other things that could be leading to it. For instance, a co-morbid medical condition. Perhaps they have a urinary tract infection. Maybe this is a delirium, that there’s some drug interaction [playing] a role. Maybe there’s a slight infection, maybe there’s a metabolic issue. So, before we go adding on another medicine, make sure that we know medically what’s also going on for that patient. Make sure our patients are not in pain before we add something else [to the] board.
HCPLive: What’s the long-term prognosis and treatment strategy for someone with agitation dementia?
McGuire: Right now, we have data that shows that we can treat this acutely. We do know that agitation, in and of itself, leads to a poor prognostic outcome for patients with dementia. We know our goal is to keep our patients in the least restrictive, most supportive environment we can. The longer we can keep them at home, surrounded [by] their family members, the better. The key is preventing those agitation episodes and treating them allows these individuals [to] have a better quality of life and to spend more time with their loved ones.
HCPLive: For patients who come in and see a psychiatrist, maybe they are dealing with depression and anxiety as a result of their stuttering. What can psychiatrists do to help?
McGuire: It’s so important because [among] people [who stutter], over 80% may struggle with social anxiety. We also know that it’s coexisting. Stuttering is a basal ganglia condition.
The key [is] to understand what treatments may make the underlying stuttering worse, and how best we can treat the underlying depression and anxiety, and attention issues. It’s going to be a comprehensive approach, understanding we’re developing, hopefully, newer treatments, that can help the core features of stuttering and then work toward cognitive behavioral and other forms of therapy, and even agents to treat the social anxiety or treat with the ADHD or the obsessive compulsive symptoms without worsening the underlying stuttering.
HCPLive: Anything else that you want to talk about this, or anything else in psychiatry that you think is important for readers?
McGuire: Well, I think what’s important for individuals that psychiatry…is [often] misunderstood because we’re dealing with really the most complex organ system, and that’s the human brain. It’s more difficult to understand than, say, the heart, the kidney.
As we learn more about neuroscience, the understanding will grow among healthcare professionals, and compassion will grow. We need better targeted treatments, and they’re going to be personalized, not one size fits all, and so the diversity of what we see in psychiatric conditions will require that diverse treatment approach as well.
Prince William and Princess Kate are extremely apprehensive about the “peace talks” between King Charles and Prince Harry, according to a close friend.
Over the weekend, The Mail reported that the Duke of Sussex’s top aides met with representatives of King Charles’s team in London in what’s being dubbed a “peace summit.”
Notably absent from the meeting was any representation from the Prince and Princess of Wales’s camp.
According to a close friend of the couple, that wasn’t an oversight.
“It’s no coincidence that William and Catherine did not have a representative at the Royal Over-Seas League,” the insider told The Mail. “They were not asked to send anyone and will be treating the talks with extreme caution.”
The meeting included King Charles’s communications secretary Tobyn Andreae and Meghan and Harry’s team, including Meredith Maines and Liam Maguire. But while it may mark a first step toward mending fences between Harry and his father, the same can’t be said for his brother.
Royal expert Hugo Vickers told The Sun that William’s “less forgiving” attitude makes reconciliation much harder given Harry’s attacks on Princess Kate over the years.
“The barbs were sent in his direction from Prince Harry, but very much directed towards Catherine,” he said.
Vickers also labelled Harry’s comments about Kate “very insulting,” including his allegations against her in his 2023 memoir, Spare.
Madeleine herself designed her bridesmaids’ baby blue corset dresses. “I had a very specific vision in mind, so that left me with one option—to make my bridesmaid dresses myself,” she says. “I have always been a hobby seamstress and have made my own looks for a few red-carpet events before, so I decided to give it a go. I have to admit they did end up taking a lot longer than I anticipated—each dress took me around 30 hours to construct and hand-embellish—but it was totally worth it. My bridesmaids looked beyond stunning and now they are the first—and hopefully not last—to ever have a bespoke Madeleine White dress.”
As golden hour settled upon Santorini, Madeleine met Andrew above the caldera. Both bride and groom admit they found themselves deeply emotional during the ceremony; Madeleine recalls trembling as she read her vows. Andrew, meanwhile, teared up as he read aloud his journal entry from the first day they met back in 2020: “I think I just met a girl who will change my life.”
Afterward, a cocktail and dinner were held on the villa’s stone balcony at sunset, which florist Bloom de Fleur and wedding planners Samkoma.world transformed into a Grecian garden. The bride changed into a backless halter gown by Berta.
After speeches from both of their fathers, as well as Andrew’s business partner Joe, the couple had a surprise for their guests: a live performance by Natasha Bedingfield, who played hits like “Pocket Full of Sunshine” and “Unwritten.” Next, vocalist Morgan St. Jean, the couple’s friend, sang their first-dance song, “Sure Thing” by Miguel. (Madeleine wore a pair of butterfly shoes by fashion student Youssef Nagib, whom she found on TikTok.)
Dancing—and plate smashing—ensued, until the night grew late enough for an after-party. “We transported everybody down to the beach a few minutes away, which we transformed into Club Fedyk,” the bride said. “We had late-night food stands dedicated to each of our pets, as well as donuts, pizza, and a poutine bar. Explaining poutine to the Greek caterers was not an easy feat, but they nailed it.”
Madeleine underwent her third outfit change of the evening: a pink chain-mail dress from Fendi and Versace’s famous Fendace collaboration. It was a nod to her TikTok creator origins. “Anyone who’s been following me for a long time knows that one of my first viral videos ever was trying to DIY the Versace spring 1999 chain-mail dress,” she says. “It felt like fate when, around a year before the wedding, I got a text from a friend in Toronto that Dawn at Rewind Couture had gotten in this one-off Fendi x Versace collab.”
Andrew, meanwhile, was hyperfocused on one aspect in particular: the music. After holding a contest on TikTok, he hired Sweatbaby, a DJ from Australia. (Although the groom couldn’t help but put some headphones on: “Andrew swore he was not going to DJ and he would take the night off, but that didn’t exactly work out,” Madeleine says, laughing. “I could not drag him off the decks at the end of the night.”) The couple and their guests stayed out until 4:30 in the morning.
As the rest of their party headed to bed, Madeleine and Andrew stayed up until the sun rose. “We got into bed on our wedding night just in complete disbelief at what an amazing time we had,” the bride says. “I couldn’t sleep for hours, just replaying all the amazing moments and trying to make sure I remembered everything.”
When asked days later about how they were feeling about the milestone, it’s clear the wedding weekend dust still hasn’t settled: “We are still reeling from just the best weekend of our lives,” Madeleine says.
The federal government on Tuesday hiked the price of petrol by Rs5.36 per litre for the next 15 days, raising it to Rs272.15.
The rate for high-speed diesel was also increased by Rs11.37 per litre, taking it to Rs284.35 per litre.
In a notification, the Finance Division stated that the new prices will take effect from July 16, 2025.
In the last fortnightly review, the government had increased the petrol price by Rs8.36 per litre to Rs266.79, and diesel by Rs10.39 per litre, to Rs272.98.
On July 15, 2025, LDI Senior Fellows Steven Joffe, MD, MPH, Holly Fernandez Lynch, JD, MBE and authors Ravi B. Parikh, MD, MPP, Martin Kurian, MD, and Ronac Mamtani, MD, MSCE delivered a memo to the FDA’s Center for Drug Evaluation and Research (CDER) and Oncology Center of Excellence (OCE) entitled Policy Recommendations for Strengthening the Accelerated Approval Program for Oncology Drugs: Response to December 2024 Draft Guidance.
In this memo, the authors assert that recent empirical research on the FDA’s Accelerated Approval (AA) program reveals critical gaps between policy intentions and real-world treatment outcomes that warrant immediate attention in finalizing the December 2024 draft guidance. The authors summarize their studies which demonstrate that current postmarketing requirement (PMR) practices are insufficiently rigorous, that prescribing patterns suggest misaligned incentives, and that substantial patient exposure to ineffective drugs occurs. They recommend the following five specific policy reforms to strengthen the Accelerated Approval program while preserving its benefits for patients with serious oncologic conditions:
Re: Policy Recommendations for Strengthening the Accelerated Approval Program for Oncology Drugs: Response to December 2024 Draft Guidance
From: Ravi B. Parikh, MD, MPP (Emory University); Martin Kurian, MD (University of Pennsylvania); Steven Joffe, MD, MPH (University of Pennsylvania); Holly Fernandez Lynch, JD, MBE (University of Pennsylvania); and Ronac Mamtani, MD, MSCE (University of Pennsylvania)
Date: July 1st, 2025
To: Center for Drug Evaluation and Research (CDER) and Oncology Center of Excellence (OCE)
Recent empirical research on the FDA’s Accelerated Approval (AA) program reveals critical gaps between policy intentions and real-world treatment outcomes that warrant immediate attention in finalizing the December 2024 draft guidance. Our studies demonstrate that current postmarketing requirement (PMR) practices are insufficiently rigorous, that prescribing patterns suggest misaligned incentives, and that substantial patient exposure to ineffective drugs occurs. We recommend five specific policy reforms to strengthen the AA program while preserving its benefits for patients with serious oncologic conditions.
The AA program represents a critical pathway for bringing promising oncology therapies to patients with serious conditions and unmet medical needs. However, recent congressional attention and growing concerns about program integrity necessitate evidence-based reforms.¹ The December 2024 draft guidance, alongside January 2025 draft guidance on confirmatory trial requirements, provides an opportunity to address systemic weaknesses while maintaining the program’s core mission of accelerating access to beneficial therapies. While the January 2025 guidance represents progress by requiring confirmatory trials to be “underway” prior to approval, our empirical research suggests that stronger safeguards are needed to ensure program integrity.
Our comprehensive analysis of 161 oncology AA indications granted between 2011-2023 reveals three fundamental problems with current AA implementation:
First, PMR statements lack the specificity needed to ensure rigorous confirmatory evidence. Among 181 PMR statements analyzed, critical design elements were frequently omitted: only 30% specified follow-up duration, 26% included enrollment targets, 24% specified requiring multicenter trials, and 13% mandated double-blinding.² These gaps translate directly into weaker confirmatory studies, as PMR studies adhered to statement specifications with near-perfect fidelity (>99%).
Second, prescribing behaviors are misaligned with the robustness of evidence. Real-world prescribing data from 63,434 patients demonstrates that oncologists respond more robustly to AA (23 percentage point increase) than to conversion to regular approval (1 percentage point increase).³ This suggests that market adoption occurs rapidly upon AA, removing commercial pressure for sponsors to complete confirmatory studies expeditiously.
Third, prescribing of ineffective therapies is prevalent. Among five withdrawn AA indications with sufficient follow-up data, 26% of eligible treatment initiations involved subsequently withdrawn therapies. Although it is inherent in AA policy, given the uncertainty of unvalidated surrogate endpoints, that some AA drugs will fail, this finding reflects thousands of patients exposed to drugs without demonstrated clinical benefit.⁴ The median time from AA to withdrawal was 46 months, creating extended exposure periods.
We recommend five specific reforms to address these identified weaknesses:
1. Implement Standardized PMR Statement Requirements
The draft guidance should mandate a comprehensive checklist of PMR statement elements, including: (a) specific enrollment targets with interim milestones, (b) defined follow-up duration sufficient to assess clinical benefit, (c) explicit randomization and blinding requirements where feasible, (d) multicenter design specifications, and (e) clinical endpoint requirements prioritizing overall survival or validated surrogates. This addresses the current variability in PMR statement specificity that undermines confirmatory study quality.⁵
2. Strengthen Pre-AA Confirmatory Study Requirements Beyond Current Guidance
FDA has undertaken important efforts to require that confirmatory trials be underway prior to AA. However, our empirical evidence suggests that merely having “enrollment initiated” at the point AA is granted is insufficient to ensure timely completion of confirmatory trials given the rapid prescribing adoption following AA. The FDA has recently responded to this by issuing recent guidance that sponsors complete all or a significant portion of enrollment prior to AA when confirmatory studies will be carried out in the approved patient population. We propose that the FDA should expect sponsors seeking AA to provide: (a) demonstrated feasible enrollment rate trajectory, (b) established data and safety monitoring procedures, and (c) interim analysis plans with prespecified stopping rules. The 23-percentage point prescribing increase immediately following AA necessitates stronger enrollment benchmarks than currently proposed.⁸
3. Expedite Post-AA Surveillance and Withdrawal Procedures
The draft guidance should establish accelerated timelines for identifying and acting upon negative confirmatory trial results. Under the recent Food and Drug Omnibus Reform Act (FDORA) of 2022, sponsors must provide progress reports on AAs every 6 months. Our data showing 26% patient exposure to withdrawn therapies underscores the urgency of robust and scrupulous tracking of emerging evidence surrounding AAs to minimize exposure periods to ineffective agents.⁶ We recommend: (a) mandatory and rapid sponsor reporting of analyses of confirmatory trials that fail to confirm effectiveness of the agent for the relevant indication, and (b) withdrawal proceedings initiated within 6 months of negative confirmatory trials, absent compelling justification for delay, unless a sponsor voluntarily withdraws. If delays continue, then tighter periods for sponsors acting on negative results will be important.
4. Enhance PMR Study Design Standards
The guidance should explicitly require that PMR studies employ design elements associated with successful regular approval conversion, including randomized, controlled designs except where explicitly justified and adequate statistical power for clinical endpoints. Our analysis showed that less rigorous PMR characteristics allowed faster completion of confirmatory trials,⁷ however, rigor is essential to resolve the uncertainty associated with AA drugs.
5. Implement Real-World Evidence Integration Requirements
Given rapid prescribing adoption following AA, FDA should require sponsors to submit real-world effectiveness analyses alongside confirmatory trial results. These analyses should: (a) assess if drug utilization patterns are consistent with approved indications, and (b) identify any emerging safety signals in clinical practice. This leverages the extensive post-AA uptake seen in our data to complement randomized trial evidence.
The AA program serves a vital public health function, but current implementation gaps undermine its integrity and potentially harm patients. Our empirical findings provide specific guidance for strengthening program requirements while preserving access to promising therapies. The December 2024 draft guidance, in conjunction with the January 2025 confirmatory trial guidance, represents a critical opportunity to implement these evidence-based reforms and restore confidence in this essential regulatory pathway.
The policy recommendations outlined above directly address the systematic weaknesses identified in our research while building upon existing statutory authorities and recent guidance developments. We urge CDER and OCE to incorporate these recommendations into the final guidance to ensure the AA program fulfills its promise of providing early access to effective cancer therapies while minimizing patient exposure to ineffective treatments.
Research Memo: Delivered to House Speaker Mike Johnson and Majority Leader John Thune
Research Memo: Delivered to House Speaker Mike Johnson and Majority Leader John Thune
Research Memo: Response to Request for Technical Assistance
Research Memo: Supplement to Response to Request for Technical Assistance
Research Memo: Delivered to Staff of U.S. Senate Committee on Finance
Testimony: Delivered to Philadelphia City Council