Radiation oncology places needles too. Epione was built for that problem.
Radiation oncology does CT guided needle work every week, and most of it is freehand. Here is where Epione could fit, what the evidence on robotic placement supports, what the label allows today, and why prostate is the long horizon rather than the start.
Why radiation oncology, in six numbers.
| 2,313Source | US radiation therapy facilities in 2020, up 16% since 2005. The US has about 5,900 active radiation oncologists, more than its roughly 4,000 vascular and interventional radiologists.Source A new department for Epione, not a slice of the one NeuWave already serves. |
| 16 to 45% | of CT guided lung fiducial placements caused a pneumothorax across published series, with chest tubes in 6% to 16%.SourceSource These markers guide SBRT, CyberKnife and proton to moving tumors. |
| 1.7 vs 4.5 mm | targeting error with a robot versus freehand in a 2026 randomized trial of CT guided biopsy, with lung complications of 4.9% versus 16.9%. A different robot and procedure, but the same needle under CT.Source |
| 32.8 vs 11.6 | months median progression free survival for CT guided interstitial HDR brachytherapy versus TACE in liver cancer, in a matched multicenter study.Source Few centers offer it, because it depends on operator skill.Source |
| 3 to 6 mm | typical error in manual seed placement.Source In one prostate study, correcting for needle tip error lowered the calculated D90, the dose covering 90% of the gland, by 12.8% on average.Source In brachytherapy, millimeters are dose. |
| 6.8 → 42% | rise in rectal spacer use among Medicare prostate radiation patients from 2017 to 2021. Radiation oncology adopts a new needle procedure fast when it helps their patients.Source |
Three places Epione could fit first, all under CT.
Fiducial markers
Gold markers let SBRT, CyberKnife and proton track tumors that move with breathing, in lung, liver and pancreas. Interventional radiology usually places them for radiation oncology, under CT, freehand. Stanford's interventional radiologists published one of the core complication series, with radiation oncology coauthors.Source The strongest transferable evidence for a robot sits here.
Interstitial HDR, liver
Catheters are placed through the skin under CT, then high dose radiation runs through them. One Munich series delivered less dose to the healthy liver than SBRT, and its authors call for close work between IR and radiation oncology.Source Local control tracks dose, and dose tracks catheter coverage. Its limit is operator skill, the problem a robot is built to address.
CT guided seed implants
Seeds for lung, pancreas and recurrent tumors, mostly in China today. With a robot assisted template, dose conformity held at 0.95 against 0.91 conventionally, with no pneumothoraces versus two.Source In phantoms, template error fell from 2.55 to 1.04 mm.Source Promising, early, and from a different system.
The common thread: the imaging is CT, the job is a straight needle to a planned point, and the result is measured in dose. That is Epione's workflow today, in a department no one at Precision IO serves. None of it uses ablation energy, so none of it overlaps with the NeuWave team.
Three things I keep separate.
Cleared today
In the US, Epione is cleared to guide needles and probes from 11 to 19 gauge during CT guided ablation in the abdomen, under general anesthesia.Source Fiducials, catheters and seeds are not in the label, and neither are the chest or pelvis. Promoting them today would be off label, so I wouldn't.
Possible future
An expanded 510(k) looks like the likely path. Perfint MAXIO and XACT ACE already hold general CT guided percutaneous clearances that could serve as predicates.SourceSource Use in the chest would likely need data on pneumothorax and breathing motion. That is regulatory's call; this role organizes the clinical input.
My hypothesis
Robotic placement may make radiation oncology's CT needle work more accurate, more consistent and safer, and could open procedures like interstitial HDR to more centers. Unproven for Epione. The plan is built to test it with four centers before anyone sells it.
Strongest transferable evidence. The chest needs a new indication.
The liver is in the label's anatomy, but brachytherapy catheters are not ablation.
Evidence mostly from China and phantoms.
Usually ultrasound or MRI guided.
Needs new imaging, a new clearance and clinical data.
Prostate: the biggest prize, on the longest clock. 333,830 new US cases this year.Source
This is where my thesis started, and it is still the largest prize. But prostate implants are guided by ultrasound or MRI, not CT, and the procedure I sold for years has been shrinking. So it comes after the CT work above, not before it.
| 69%Source | of prostate cancers are localized at diagnosis, and incidence has risen about 3% a year since 2014. |
| 47%Source | drop in Medicare LDR seed implants from 2013 to 2022, from 4,374 to 2,296. HDR held steadier, near 3,000 a year. All payer volume is higher, but no reliable national count exists. |
| 1.3%Source | of men with prostate cancer received focal therapy from 2010 to 2023. Small, but the mix is shifting fast toward laser and HIFU. |
| +39%Source | growth in EDAP's 2025 HIFU revenue, to about 33 million euros. Profound reports 78 TULSA systems installed. Company revenues are a harder signal than analyst sizing.Source |
Each grid is 100 patients with localized disease. Filled holes received brachytherapy. National Cancer Data Base, Martin et al.Source Drivers included robotic prostatectomy, IMRT and SBRT, reimbursement, and fewer physicians trained in the technique. Analyst estimates for prostate ablation devices include BPH and overstate the cancer opportunity, so I leave them out.
How it is done today, and where placement gets hard.
| Procedure | Guidance and owner | Where placement is hard | Who sells it |
|---|---|---|---|
| Brachytherapy, LDR and HDR | TRUSTemplateTransperineal needles through a template. Radiation oncology with urology. | Mean seed misplacement of 4.9 mm in one study, from needle deflection, edema and prostate motion.Source Pubic arch interference blocked needle paths in about half of patients in another series.Source | Radiation oncology vendors |
| Needle based focal ablation | TRUSMRI fusionCryo, IRE and laser through the perineum. Urology. | IRE needs several needles placed parallel at set spacing around the tumor. That geometry is what planning robots do. | AngioDynamics NanoKnifeBoston Scientific cryoAvenda laser |
| Nonneedle focal therapy | TRUSMRIHIFU through the rectum, TULSA through the urethra. Urology. | No percutaneous needles, so a placement robot adds little. | EDAP Focal OneSonaCareProfound TULSA |
| Salvage ablation of recurrence | CTSometimes done by interventional radiology. | Small series only, such as 18 CT guided cryoablations with 95% technical success.Source The closest prostate work to today's Epione workflow. | Cryo vendors |
| Spacers and fiducials | TRUSTransperineal needle placement before radiation. | Precise but low value per case, and already routine under ultrasound. | Boston Scientific SpaceOAR |
TRUS: transrectal ultrasound. Filled chip: the imaging Epione uses today.
A robot already exists here: Biobot's iSR'obot Mona Lisa is FDA cleared for transperineal needle positioning with MRI and ultrasound fusion.Source Research robots have reached 1 to 2 mm accuracy in phantoms and small series, but after twenty years no robotic brachytherapy system is in routine use, and I found no comparative outcome data.Source
What Precision IO would have to solve first.
| Barrier | Epione today | Prostate needs |
|---|---|---|
| Imaging | CTCBCT, EUCone beam CT in Europe only | TRUSMRI fusionCT shows prostate tumors poorly. |
| Clearance | Abdominal ablation in the US | A new indication, backed by performance and clinical data |
| Positioning and access | Patient on the CT table, general anesthesia in the label | Transperineal access in lithotomy, often in an OR or procedure room |
| Buyer | Interventional radiology | Urology and radiation oncology, with referral politics between them |
| Payment | Established ablation codes; robotic guidance is not paid separately | Brachytherapy and HIFU have codes, but many payers still restrict focal therapySource |
| Evidence | Accuracy series in liver and kidney | Data showing placement improvements change dosimetry, workflow or results |
How the research changed my thesis.
What I started with
Robotic guidance could make prostate brachytherapy needle placement more accurate and consistent, and win back a procedure that lost ground partly because fewer physicians are trained to do it well.
What I believe now
Start where Epione's imaging already fits: CT guided fiducials, interstitial catheters and seeds. Treat radiation oncology as the customer, not a referral source, and keep the work separate from NeuWave. Prostate stays in view as the long horizon, with MD Anderson's MRI guided program as the place to learn what it would take.
| What I would test | How |
|---|---|
| Is placement a real pain? | Structured interviews with radiation oncologists, physicists and the IR teams who place their fiducials at Ohio State, Vanderbilt, Stanford and MD Anderson, using the same questions, in the first 60 days. |
| How big is it? | No national count of CT guided fiducial or interstitial procedures exists. Build one from the four centers' volumes and Medicare claims for fiducial placement (CPT 32553 and 49411), with marketing. |
| Is the product path real? | Ask regulatory and R&D what an expanded indication would take: anatomy, instrument types, predicates, and whether bench data or a clinical study is needed. |
| What would end it | Radiation oncologists who see no placement problem, or a regulatory path longer than three years for the first use. Then I would narrow to the single use with the best evidence and park the rest. |
In the role: convene leading radiation oncologists through medical affairs, capture each workflow step by step, and hand R&D and regulatory a clear answer. I won't claim outcomes nobody has proven.
Five radiation oncology uses, ranked.
Scored 1 to 5 on what I can see from public information, where 5 is best for Precision IO. These are my starting estimates, to be rescored with the four centers in month one. Change the weights to see how the order holds.
| Rank | Use | Size | Clinical fit | Technical fit | Regulatory ease | Access | Competitive room | Time to revenue | Score |
|---|
Gynecologic: only 15% of cervical cancer brachytherapy patients get interstitial implants, partly because placement is hard.Source Robotic catheter placement is already an NIH funded research topic.Source Set aside: SBRT itself, which needs no needle, and rectal spacers, which are routine under ultrasound.