Keeping up with oncology drug development is a full-time job. You already have one.

Torralo is your own oncology intelligence desk, inside the AI you already use. Ask a question about a drug, a target or a disease, and get the landscape you would otherwise spend a week assembling.

Your AI assistantTorralo connected

What CAR-T therapies are in development for multiple myeloma?

Under FDA review

  • Anito-cel (anitocabtagene autoleucel), a BCMA CAR-T. Arcellx developed it with Kite, and Gilead bought Arcellx in April 2026.2 Gilead says the FDA plans to decide by 23 December 2026.1Anito-cel, Gilead (from Arcellx and Kite), BCMA CAR-T.2 FDA decision due 23 December 2026.1

Phase 3

  • Arlo-cel (arlocabtagene autoleucel), Bristol Myers Squibb, a GPRC5D CAR-T. Positive registrational phase 2 results, September 2026; phase 3 against standard regimens under way.Arlo-cel, Bristol Myers Squibb, GPRC5D CAR-T. Positive phase 2, September 2026; phase 3 under way.3,4
  • AZD0120 (GC012F), AstraZeneca, a CAR-T against BCMA and CD19. AstraZeneca is testing it in relapsed myeloma, and against stem cell transplant in newly diagnosed myeloma.AZD0120, AstraZeneca, BCMA and CD19 CAR-T. Relapsed and newly diagnosed myeloma.5,6

Phase 1/2 (1 of 12)

  • CT071, CARsgen, a GPRC5D CAR-T, in relapsed myeloma.7

Already approved: Abecma and Carvykti in the US, and Fucaso and zevor-cel in China.Approved: Abecma and Carvykti (US), Fucaso and zevor-cel (China).8,9,10,11

11 sources · As of 11 November 2026
  1. Gilead and Kite, ASH 2026 data release, 4 November 2026
  2. Gilead, completion of the Arcellx acquisition, 27 April 2026 (at that time Gilead gave 2 February 2027 as the FDA date)
  3. Bristol Myers Squibb, QUINTESSENTIAL phase 2 topline results, 8 September 2026
  4. ClinicalTrials.gov NCT06615479, arlo-cel phase 3 (QUINTESSENTIAL-2)
  5. ClinicalTrials.gov NCT07391657, AZD0120 phase 3 in relapsed myeloma (DURGA-4)
  6. ClinicalTrials.gov NCT07735637, AZD0120 phase 3 versus transplant (DURGA-6)
  7. ClinicalTrials.gov NCT06333509, CT071 phase 1/2
  8. FDA approval letter, Abecma, 26 March 2021
  9. FDA approval letter, Carvykti, 28 February 2022
  10. IASO Bio and Innovent, China approval of Fucaso, 3 July 2023
  11. CARsgen, China approval of zevor-cel, 1 March 2024

How does it work?

  1. Connect.

    You add Torralo (an MCP server) to the AI assistant you already use (ChatGPT, Claude, Copilot, and more). It takes a minute.

  2. Ask.

    You ask your question in plain English. Your assistant uses Torralo on its own. If it needs a nudge, tell it: “Use Torralo.”

  3. Build.

    You ask your assistant to turn the answer into slides, infographics or spreadsheets, and to update them whenever you need.

What data does Torralo have?

Every answer cites its source: company pipelines and filings, regulators, press releases and trial registries.

  • Competitive landscapes

    8,143 cancer drugs and 41,957 trials, by target, biomarker or disease, with owner and stage

  • Catalyst calendar

    3,370 upcoming readouts, filings and regulatory decisions

  • Company pipelines

    3,204 companies with 100% pipeline coverage of the top 200 companies

  • Approval histories

    2,689 approvals by indication, date and patient population, covering every cancer drug approved in the US and EU

  • Deal terms

    654 licensing deals, acquisitions and partnerships, with the upfronts and milestones companies disclosed

  • Pivotal trial results

    32,052 efficacy results from 4,789 trials: response rates, survival and safety, side by side

How does information get into Torralo?

A team of AI agents does the work, from the first fetch to the filed record.

A busy two-storey robot workshop. Sheets ride a conveyor along the mezzanine, where robots pick them up, read them and put them back, then drop down a tube onto a heap; robots walk them to bins, pour the bins onto a belt, scan and stamp each sheet, and a writer reads each one and types a new gold sheet with its quote, filing the original in a vault and zapping rejects. Checkers send gold sheets back to be fixed or on up a tube to be bound into binders on a rotating carousel; a fetcher wheels a binder to a lookup machine, where questions shoot in, a scanner reads each one against the binder, and answers shoot back out. ““““““AAAAA? LABELS TRIALS FILINGS PRESS
  1. Collect.

    Agents gather data from companies, regulators, and registries around the globe.

  2. Sort.

    Documents are separated and sorted by type.

  3. Analyze.

    Important facts are extracted and summarized.

  4. Check.

    Facts are compared again to original sources and corrected if errors have emerged.

  5. File.

    The most important data is organized for quick retrieval.

  6. Answer.

    Your AI assistant asks a question. Torralo looks it up in those records and sends back the answer with its sources.

Who is behind all this?

Hi, I’m Mike Italia. I help healthcare and life sciences companies build AI products. My client work requires me to stay current on oncology. I got tired of piecing the picture together from registries, press releases and filings, so I built Torralo.

Products that do this already exist. I couldn’t justify the price, and neither can most startups, small companies and independent consultants. The price is high because vendors pay rooms full of analysts to keep the data current. AI changes that math. I run Torralo myself, with a team of AI agents, and I’m pricing it for the people large vendors leave out.

Get an invitation to the beta

Leave your email. I’ll contact you when the beta is ready and invite you to take part.