Former Health and Human Services chief Xavier Becerramay be surgingin the competition for California governor, but his ties toan alleged fraud caseare still dogging him.
A longtime aide for Becerra, who’s considered a Democratic frontrunner, pled guilty to charges that he helped illegally funnel campaign cash towards a no-show job for his wife — and the governor hopeful is now facing uncomfortable questions about the bombshell corruption scandal.
According to the feds, Gov. Gavin Newsom’s former chief of staff Dana Williamson plotted with lobbyist Greg Campbell and longtime Becerra aide Sean McCluskie to siphon $225,000 from Becerra’s dormant campaign account for personal use between February 2022 and September 2024.
While McCluskie was Becerra’s chief of staff, Williamson — who also worked on Becerra’s 2018 campaign for attorney general — allegedly funneled $10,000 per month. The funds were routed through multiple business entities and falsely labeled as pay for a “no-show” job that didn’t exist.
Williamson also faces other charges, and Newsom had said he placed her on leave when made aware of a probe. She pleaded not guilty, while Campbell and McCluskie pleaded guilty.
Becerra has repeatedly denied knowledge of the scheme, and has not been accused of wrongdoing.
But his allies’ alleged misdeeds have put an unflattering spotlight on the longtime California pol.
Former Los Angeles Mayor Antonio Villaraigosa, who is also running for governor, blasted Becerra online Wednesday morning for what he called a flip-flop on what Becerra knew around the scandal.
“Same politician. Same illegal payments,” said Villaraigosa, who is also a Democrat. “But five months later Xavier Becerra tells a different story.”
The former mayor shared two clips, put on repeat and dramatically slowed down, where Becerra tells one television station in November that “I was aware of the payments being made.” Then he tells another journalist in April that “I didn’t oversee that part.”
In the November clip shared by Villaraigosa, Becerra said he “had authorized” the payments when approached by advisors and “saw” the payments.
Dana Williamson, a former top aide to Gov. Gavin Newsom, center, leaves the courthouse.AP
Becerra adviser Michael Bustamante flatly dismissed the attack.
“I’m sure we’re going to see plenty of tweets from Antonio over the next few days as he sits on the couch while the rest of the legitimate candidates are debating on stage,” Bustamante said.
“There is no substance. This is all contrived from a candidate who is probably moments away from exiting,” he added.
A full look at the November interview shows that Becerra said he was misled as to the nature of the payments, which he was told was for account management services. He believed them, he said.
“If you take a look at what campaigns spend, you spend— you have to pay quite a bit of money for compliance purposes, for oversight, for legal fees, for basic management, for filing of your financial disclosure statements, for filing of your tax returns. It’s a lot of activity,” Becerra said.
He was busy working in the Biden administration running the Health and Human Services agency, so “I was told its going to cost us about $10,000— I was told that’s the rate I would have to pay to get someone who could manage that and make sure that I don’t have to worry about it.”
Antonio Villaraigosa speaks during a gubernatorial candidate forum in Sacramento, Calif.AP
In the new April interview, he clarified he “had no sight on that” as he was busy with HHS. When asked what it says about his judgment to trust McCluskie, he acknowledged, “People make mistakes.”
The California Post reached out to the Becerra campaign for comment.
Becerra had been trailing in the single digits in polls, but after former congressman Eric Swalwell left the race from sexual assault allegations, he has surged to become one of the leading Democrats.
Villaraigosa still polls near the bottom of the field. He is a frequent thorn in Becerra’s side; the former mayor even launched attack ads on Becerra before his rise in the polls.
United States Secretary of the Navy John Phelan is stepping down "effective immediately," chief Pentagon spokesman Sean Parnell said on X.
US Navy Under Secretary Hung Cao will become acting secretary, Parnell added, offering no reason for Phelan's departure, which comes amid the ongoing US naval blockade against Iran.
Phelan, who was confirmed for the role in March of last year, is the second high-ranking US military official to leave this month, after the Pentagon announced the early retirement of US Army Chief of Staff General Randy George at the start of April.
The frontier of RNA science is very nonmessenger. A decade ago, the noncoding RNA field was likened to dumpster diving.
Hani Goodarzi
Obscured in the bottom corner of the dumpster were what have come to be known as orphan noncoding RNA, or oncRNA. It’s still unclear the purpose of this new type of genetic material, which was discovered and named in Palo Alto-based researcher Hani Goodarzi’s lab in 2018.
Now, his team’s latest findings push further forward the notion that oncRNA-based liquid biopsies may someday play a pivotal — and, excitingly, multipurpose — role in precision oncology and beyond.
Presented in Cell Reports Medicine, Goodarzi’s team showed there are 260,968 different oncRNA that combine with predictable patterns likened to digital barcodes so specific that they suggest entirely new tumor subtypes across a range of cancers whose cells secrete oncRNA into the blood. A proof-of-concept phase of the research linked oncRNA with breast cancer metastasis.
The extensive nature of the paper and its pan-cancer analyses are buzzworthy, onlookers say.
“It’s not just a risk classifier. It’s not just a diagnostic molecule. It’s not just a tool that will help us subtype tumors. It really is hitting on a lot of different areas, and I think that’s what might be garnering the buzz around these orphan noncoding RNAs,” said Yitz Goldstein, MD, section chief of molecular diagnostics in the Montefiore Medical Center in Bronx, New York. He wasn’t involved in the study.
Goodarzi admitted every bit of his team’s years of work was piled into the single paper because most have left to join the startup company he co-founded, Exai Bio, and “the overhead of multiple papers wasn’t an option.”
The RNA Bet
Why Goodarzi is even in the middle of the wild west of RNA is serendipitous, he acknowledged. He started his lab in 2016 and wanted to investigate whether cancer cells can engineer their own pathways of gene expression control.
“To be honest, we didn’t really set out to solve a molecular diagnostic problem. It was actually one of these high-risk, high-reward kind of bets,” said Goodarzi, a PhD multidisciplinary scientist with expertise in AI and machine learning and in cancer systems and RNA biology. He is a co-investigator at ARC Institute and an associate professor of biochemistry and biophysics at the University of California San Francisco (UCSF).
Goodarzi described his initial research question as: Is cancer really a dysregulation of normal cells — an evolutionary disease — or do cancer cells go “beyond what is available to them in normal cells, and instead do just entirely new things — new molecular mechanisms?”
Ash Alizadeh, MD, PhD
The 2018 identification of oncRNA garnered a lot of skepticism for its translational potential. Stanford professor of medicine and liquid biopsy researcher Ash Alizadeh, MD, PhD, knows how arduous the translational path is — his lab’s work has led to new technologies that made it to market. He’s followed Goodarzi’s work and said his skepticism of oncRNA persists but is much lower after the latest paper.
“I am a little bit more convinced than I was a decade ago that these things can be functional,” said Alizadeh, noting two main aspects driving oncRNA’s clinical prospects. “The number of these oncRNAs across cancer, at a quarter million, seems like a lot. That’s a big number to have your pick from. The second is the diversity and the pattern of their variation across cancers in a way that can serve as a barcode of the identity of a lineage — a fingerprint of the tumor type as molecules that you can count digitally and do cool things with. Those are the big surprises for me that I think are likely to be important.”
The ‘Aha Moment’
The foundational oncRNA work was primarily done in breast cancer because of Goodarzi’s relationship with the I-SPY breast cancer researchers at UCSF who are attempting to personalize breast cancer treatment by finding which cancer types respond to which treatments. So when it came time to expand their understanding of oncRNA, the team at first considered looking at colon cancer because they also work in that space.
“Then we thought, maybe we’ll just do it across the board,” Goodarzi recalled, and applied their efforts to the entire dataset from the Cancer Genome Atlas.
First author and bioinformatician research assistant Jeff Wang created a binary map of oncRNAs that quickly demonstrated two important things. First, there are tens of thousands of them. And second, it was visually apparent there are cancer-specific patterns.
“That was another kind of ‘aha’ moment that you showed that you can actually train good classifiers simply based on which oncRNAs are present in a sample, not even at the expression level, but just presence or absence, and tell which tumor it comes from,” Goodarzi said.
At that moment, he paused and asked “what it really meant to have these cancer-emergent RNAs? So where do they actually come from?”
So they traced them.
About 60% come from longer RNA, while the remaining 40% trace back to cancer tissue but not surrounding healthy tissue, suggesting that — as they explained in the paper — “oncRNA biogenesis may arise from cancer-emergent transcription events in regions with increased chromatin accessibility.”
“They come from the same regulatory processes now rewired and broken in the context of cancer, but they are part of the gene regulatory network of the cell,” Goodarzi said. “And we showed, for example, a large part of them very clearly come from dysregulations in the epigenome. So parts of the chromatin that is usually closed opens up and you get new transcriptional events in these regions and you get these now fragments of RNA that we have defined as oncRNAs. We are capturing this gene regulatory network of the cell.”
This is not, however, what’s known in the traditional sense of reprogramming of the cancer genome during disease, which “usually manifests itself as gene expression changes, which we have loved and studied to death for like three decades and everywhere,” Goodarzi said.
“Now it turns out that there’s this flip side of the coin, where these regulations are giving rise to these new RNAs that are, instead of going up and down, they’re appearing as a zero-one digital signal,” he said.
Still…What Do oncRNA Do?
The majority of what we know about cancer biology is based on the understanding that proteins execute functions. Still, no one knows what oncRNA do.
Bypassing the existential question is part of what propelled the discovery.
Nadya Dimitrova, PhD
Long noncoding RNA researcher Nadya Dimitrova, PhD, called the Goodarzi team’s work “really creative” because at the time most other scientists were studying RNA with known functions.
“Everybody else was looking at the big shiny thing, and they discovered an even bigger trove of data that everybody was just glossing over and ignoring,” said Dimitrova, associate professor in the Department of Molecular, Cellular, and Developmental Biology at Yale School of Medicine, New Haven, Connecticut.
To learn that oncRNA can be cancer-specific, defining tissue of origin and differentiating between cancer subtypes “are very important findings,” she said, although not entirely surprising because other noncoding RNA work has come to similar conclusions and yet “physicians, physician scientists, and most biologists have ignored noncoding RNAs.”
The Goodarzi work reinforces “that these noncoding RNAs, even if we don’t know where they come from, even if we don’t know what they do, we know that they’re actually a more accurate representation of the cancer,” Dimitrova said.
In this protein-centric world, the pressure is on to perform good science and drive the field of noncoding RNA forward now that the potential has been well-defined, she added.
Clinical labs are ready and waiting. Goldstein said he’s keeping tabs on scientific papers like this because he knows breakthroughs are imminent based on recent hematologic advances. The currently used biomarker panel at Montfiore can guide therapies in less than a day for many cancers, and a recently adopted panel for suspected acute leukemia can guide targeted therapy in two days’ time, compared to 7 to 10 days for results at other labs. The solid tumor panel has remained unchanged, though, for several years.
“What this paper really highlights is that the more we learn, the more we realize the less we know. And that there’s an ever-growing ‘-ome,’” Goldstein said. “There’s the proteome. There’s the methylome. There’s now an RNA-ome — a transcriptome. And so there’s a never-ending -ome of learning here, of new molecules. Many of those will not make their way to the bedside, but the ones that do are already having an impact in patient care. And we can really be excited about what’s to come.”
Alizadeh disclosed that he is the co-founder of the company Resero Bio. Goodarzi disclosed that he is a co-founder and board of directors member of Exai Bio. Disclosure information for study authors is available in the original study publication.
A tau blood test may be able to predict Alzheimer’s disease (AD) years before brain scans reveal pathology or symptoms develop.
In a longitudinal cohort of cognitively healthy older adults, investigators found elevated baseline levels of plasma phosphorylated tau 217 (pTau217) predicted faster accumulation of AD pathology on imaging.
“What stood out in our study is that even when amyloid scans appear normal in the clinic, the pTau217 biomarker can identify individuals who later become amyloid-positive,” study investigator Hyun-Sik Yang, MD, neurologist at Neuroscience Institute, Mass General Brigham, in Boston, said in a statement.
“We used to think that PET scan detection was the earliest sign of Alzheimer’s disease progression, revealing amyloid accumulation in the brain 10-20 years before symptoms appear. But now we are seeing that pTau217 can be detected years earlier, well before clear abnormalities appear on amyloid PET scans,” Yang said.
Equally important, Yang added, is that the data show that individuals with low baseline pTau217 levels are likely to stay amyloid-negative for several years.
The study was published online on April 14 in Nature Communications.
Early Predictive Power
To test the early predictive power of plasma pTau217, the researchers followed 317 cognitively unimpaired older adults from the Harvard Aging Brain Study for an average of 8 years, with repeated blood tests, amyloid and tau PET scans, and cognitive testing.
Baseline plasma %pTau217 was strongly associated with baseline brain amyloid-beta burden measured by amyloid-beta PET. The biomarker classified amyloid-beta status with high accuracy (area under the curve, 0.94), with sensitivity and specificity of 0.89 and 0.90, respectively, at the 4.2% threshold recommended in the assay used for the study.
Higher baseline plasma %pTau217 was associated with a longitudinal increase in brain amyloid-beta measured by PET. Even among individuals who were amyloid-negative at baseline, higher pTau217 predicted future increases; those in the highest tertile were less likely to remain amyloid-negative at 6 years than those in the lowest tertile (72% vs 98%).
Notably, the investigators reported, the trajectory of %pTau217 and amyloid-beta Centiloid over time followed a “sigmoid curve, suggesting that plasma %pTau217 increases before substantial cortical amyloid-beta accumulation.”
Participants who were amyloid-beta-negative and had very low pTau217 (below ~2.6% in this cohort) rarely became amyloid-beta-positive, suggesting this subgroup may be low risk and might not require amyloid-beta PET until pTau217 rises, the researchers said.
Higher baseline plasma %pTau217 also predicted tau accumulation in the brain — even when amyloid levels were still low, suggesting it captures very early disease processes.
In terms of cognition, across the full cohort, higher plasma pTau217 was associated with faster decline on a composite cognitive score. However, this relationship was largely driven by individuals who already had elevated amyloid; among amyloid-negative participants, pTau217 did not significantly predict cognitive decline over the follow-up period.
Further analysis suggested a cascade in which elevated baseline pTau217 leads to amyloid accumulation, which then drives tau buildup, ultimately resulting in cognitive decline.
Not Ready for Prime Time
While promising, the results do not yet support widespread clinical screening using pTau217 alone without confirmatory testing, the authors said.
“For now, pTau217 as a screening test, would not be advised because we don’t know the clinical benefit of doing that and because there’s no intervention that has been shown to be effective in this early stage,” Yang told Medscape Medical News.
“However, in research, if we screen people, and they have very low pTau217, the question is — do we even need to repeat an amyloid PET or do we even need to enroll them in Alzheimer’s studies? That’s something that we need to look into,” Yang said.
Reached for comment, Maria C. Carrillo, PhD, chief science officer and medical affairs lead at the Alzheimer’s Association, said this new study “adds to growing scientific evidence” that the plasma biomarker pTau217 is a strong indicator of AD disease pathology buildup in the brain.
What’s novel here, said Carrillo, is that plasma pTau217 may detect initial buildup of AD-related brain changes before they can be detected by current “gold standard” tests, such as PET scans.
“As the Alzheimer’s Association and all other stakeholders in this field anticipate results of clinical trials of disease modifying treatments in people with preclinical Alzheimer’s, the urgency for accurate, dependable, and accessible diagnosis at this earliest stage of the disease is high,” Carrillo told Medscape Medical News.
To help dementia specialists navigate the fast-changing landscape of blood testing, the Alzheimer’s Association recently released its first-ever clinical practice guideline for use of AD blood tests in specialty care settings.
“The guideline gives specialists evidence-based advice on who should get tested, which test to use, and how to interpret the results. It will be updated as new research emerges,” Carrillo said.
The study was supported by grants from the National Institute on Aging and the Shelby Cullom Davis Charitable Fund. Disclosures for study authors are available in the original study publication. Carrillo had no relevant disclosures.
Optum Rx — the pharmacy benefit manager for UnitedHealth Group — claims its “PreCheck” prior authorization tool not only cuts prescription approval times but also reduces denials and appeals.
UnitedHealth Group gave an update on the tool in an April 21 earnings call. Optum CEO Patrick Conway, MD, said denials due to missing information dropped by 68% and appeals were down 88%, thanks to PreCheck. He said PreCheck has been “easing interactions for clients, members and providers.”
Dr. Conway reaffirmed that PreCheck has axed prescription approval time from eight hours to fewer than 30 seconds.
Optum Rx announced an expansion of PreCheck in November, alongside its decision to eliminate reauthorization requirements for 40 medications. In the November release, UnitedHealth Group said, as of this year, the PreCheck platform covers more than 45 medications and is leveraged across 20 health systems.
Insurers have been rolling back prior authorizations since an industrywide pledge last summer, which included UnitedHealthcare. The insurer’s stated goal is to cut prior authorizations by at least 30% by the end of 2026.
Additionally, Optum Rx onboarded more than 800 clients but lowered call center volume by one-quarter through elevated digital and AI-enabled self-service. Member satisfaction surpassed 95%, according to Dr. Conway.
This year, UnitedHealth committed to a $1.5 billion AI spend as it continues to explore use cases, particularly as the tech-enabled services business Optum Insight transitions to an AI-first model.
A new class of contenders is entering theAI arms racebetween payers and health systems.
Several AI startups are now using large language models to fight insurance denials on behalf of patients with early results that are turning heads.
Claimable — founded by British physician and entrepreneur Warris Bokhari, MD, in 2023 — uses an AI model trained on insurance law, medical literature and legal precedents to generate and file appeal letters on behalf of patients. For $50, users submit their case details on the company’s website. Claimable sends the resulting letter to the insurer’s appeals department and, in some cases, to executives, politicians and journalists to add pressure, Bloomberg reported April 22.
About 4,000 patients have used the platform, with about 3 in 4 seeing their denials reversed. In one case, a denial was overturned within a day after Claimable escalated it to the patient’s senator, governor, the insurer’s CEO and investor Mark Cuban, who emailed the employer’s CEO and posted publicly about the situation on LinkedIn.
Claimable is also moving beyond the direct-to-consumer model. The company has signed deals with four drugmakers and is negotiating with health systems to handle appeals on behalf of their patients who have had treatments denied. In addition, the company is exploring a litigation arm to pursue class-action lawsuits when it identifies patterns of allegedly wrongful denials across payers.
Sheer Health, another startup, takes a broader approach, handling the entire appeals process on behalf of patients for $40 a month or a percentage of money recovered, according to CBS News. Users submit their bill via photo upload and the company manages all insurer communication and paperwork.
“Our goal is for people to never have to deal with their health insurance again,” co-founder Ben Howard told the publication.
The startups are entering a market shaped by years of escalating friction between payers and hospital revenue cycle teams over claims and prior authorization. AI is accelerating that dynamic on both sides. As automation makes it faster and easier to deny claims, it is also making it faster and easier to fight them.
Whether patient-facing tools like Claimable and Sheer Health can meaningfully shift that balance remains to be seen, but the gap they are targeting is real. Claimable estimates insurers deny 850 million claims annually, extrapolated from federal data, though the company did not disclose which dataset or how the figure was calculated. Consumers appealed fewer than 1% of denied claims in 2023, and when they did, insurers upheld their original decision 56% of the time, KFF data shows.
Lawmakers are also starting to take notice. Arizona passed a law that takes effect this June requiring licensed medical providers to have the final say on coverage denials, and Louisiana is advancing similar legislation that would require human clinical review before any denial is issued.
The American Medical Association iscallingon federal lawmakers to enact safeguards on AI chatbots, particularly when it comes to protecting patients’ mental health.
With the rapid rise of people using chatbots for behavioral and other health-related issues, the AMA wrote April 22 to the co-chairs of the Congressional AI and Digital Health caucuses and the Senate AI Caucus, urging stronger regulation of the technology.
“AI-enabled tools may help expand access to mental health resources and support innovation in healthcare delivery, but they lack consistent safeguards against serious risks, including emotional dependency, misinformation, and inadequate crisis response,” AMA CEO John Whyte, MD, stated in a news release. “With thoughtful oversight and accountability, policymakers can support innovation and ensure technologies prioritize patient safety, strengthen public trust, and responsibly complement — not replace — clinical care.”
The AMA’s recommendations include requiring chatbots to clearly disclose that users are interacting with AI, prohibiting them from presenting themselves as licensed clinicians, banning them from diagnosing or treating mental health conditions without regulatory due diligence, clarifying when AI solutions qualify as medical devices, and mandating strict data protection standards.
