This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here.

Co-written by Dr Sanjay Parekh

Optellum, a UK-based AI vendor focused on lung nodule malignancy prediction, has recently announced that the Centers for Medicare and Medicaid Services (CMS) has established a national payment rate for the vendor’s technology, under the New Technology Ambulatory Payment Classification.

The move means that health insurance claims will be able to be paid out for individuals enrolled in Medicare. This will help facilitate the use of Optellum’s tool and financially incentivise providers to adopt it, but could it also point to medical imaging AI’s transition into the mainstream?

The Signify View

Much attention and many column inches are given over to medical imaging AI’s most prominent vendors. Interest is, quite rightly, usually directed at those vendors that have secured vast amounts of funding or announced grand partnerships. However, these aren’t the only companies worth watching, with smaller, less well-known vendors sometimes providing the most impactful and market-shaping stories.

Such is the case with Optellum. It is a relatively small vendor, having only raised a small amount of funding since its inception in 2015 (we estimate < $10 million), an amount that pales into near insignificance when compared with some other recent funding activities, propelling AI vendors to total funding hauls of $250m and beyond. The vendor was also quiet in terms of regulatory approvals until March 2021, when its solution received approval from the US-FDA. So far, so normal.

Where interest may start to be piqued, however, is when the nature of Optellum’s tool is looked at more closely. Many vendors offer lung cancer detection and quantification AI tools for CT imaging, including large medical imaging vendors such as Siemens (AI-Rad Companion Chest CT) and Philips (DynaCAD Lung), and smaller specialist vendors such as Riverain Technologies, Aidence, and Infervision.  But, where Optellum sets itself apart is its focus on malignancy probability, with the vendor’s solution able to predict the likely malignancy of a lung nodule. Instead of simply assisting or providing additional information to a radiologist, the tool should enable a radiologist to work more efficiently and effectively by prioritising those that are more likely to have malignant, rather than benign nodules.

Popular Algorithms

Lung cancer is a pathology for which early detection can have a particularly significant impact. A lack of screening programmes means that often, early detection isn’t happening. Optellum could be useful in direct screening programmes, but, more significantly in the near term, the solution could be used for incidental screening, being used on every chest CT conducted, and drawing attention to those that look likely to be malignant. Optellum’s use transcends being applied solely to individual patients and is also applicable in a population health context.

Most of the recent CPT codes that have come into force (which itself is a significant step forward in allowing the use of certain tools to be tracked) were for quantitative assessment. Some vendors may leverage these CPT codes for their solutions that improve workflow efficiency, but these solutions may not necessarily have the same population health benefits.

Optellum has successfully demonstrated that its Virtual Nodule Clinic solution qualifies for 0721T (or 0722T) for quantitative CT tissue characterisation as it measures the likelihood of lung cancer on CT imaging. Its ability to predict lung cancer, a capability lacked by many other AI vendors with lung nodule solutions, is the reason why Optellum’s lung cancer prediction technology received reimbursement, when other solutions which have recently been provisioned Category III CPT codes have not.

For instance, the very same CPT code for CT tissue characterisation, could also be applied to coronary artery calcium (CAC) scoring from vendors such as Cleerly, or Nanox.ai (formerly Zebra Medical Vision). However, CAC scoring could be indicative of coronary artery disease, but it is not directly linked to that pathology. Optellum’s solution, conversely, directly assesses the probability of the patient having the pathology itself, rather than a biomarker of disease that indicates a patient may have or is at risk of said pathology. By focusing on driving improvements in a care pathway rather than incremental increases in diagnosis efficiency or accuracy, Optellum’s solution will also be favoured more by value-based care health systems, such as single payor markets, or government-driven programmes such as Medicaid. Moreover, it will also certainly raise the eyebrows of private payors as a potential disruptive and competitive influence on traditional care models.

Further, the fact that Optellum’s solution has already received reimbursement, despite it only being granted a Category III CPT code (typically, with potential for, but no guarantee of reimbursement) adds further credibility to its value and perceived impact it will confer. Breast cancer CAD solutions (computer-vision and basic machine learning) were the first wave of software tools to receive reimbursement. Providers used these tools to claim the reimbursement despite many of them not performing as well as expected or not providing real clinical value, by returning high rates of false positives, for example. The provision of Category III CPT codes without reimbursement helps prevent this, with providers only using the tools that they see value in, regardless of whether it is reimbursed or not. Adoption and usage of these tools will, ultimately, result in the Category III CPT codes being upgraded to Category I, which is assigned obligatory reimbursement. Given that Optellum’s solution already receives reimbursement, suggests it has overcome the “value” threshold, with the CMS only offering reimbursement on tools it is confident provides an advantage.

Joining the Ranks

Optellum’s reimbursement also represents a positive change for the broader medical imaging AI industry.  In the first instance, the reimbursement shows what other vendors must demonstrate to secure codes and reimbursement for their own solutions. Those which already qualify for the new wave of CPT codes will look to demonstrate, as Optellum has done, the same or equivalent value, while those which do not yet qualify for CPT codes will also use Optellum’s progress to push for the CPT codes, which will represent the first steps on the road to reimbursement. A clear example of this is ScreenPoint Medical, which offers quantitative mammography tissue characterisation; a tool very comparable to that of Optellum, albeit in a different modality for a different cancer. At present there is no CPT code that ScreenPoint can utilise, but Optellum’s success will give ScreenPoint greater incentive to pursue the same, especially as it looks to enter the highly competitive US market.

The awarding of reimbursement, and particularly the pace at which it was given, also offers insight into the type of population health tools that the CMS will look to reimburse. Such a move therefore makes it look more challenging for tools that are only capable of ‘detect and triage’ to make the same progress because they don’t offer the same value proposition for clinicians and providers or the same benefit to patients.

Population health tools in contrast, whether used to pick up incidental findings or in screening programmes, can offer earlier detection leading to significant improvements longer-term patient health and ultimately better outcomes at lower costs. The CMS is effectively looking to guide the medical imaging AI market through its reimbursement and in effect remove one of the biggest barriers to AI adoption, allowing these population health solutions to prosper.

Risk and Reward

In removing, or at least mitigating the cost of certain AI tools, the CMS is significantly reducing the financial risk a provider takes in adopting an AI solution. Instead of providers having to derive a return on is investment at some indeterminate point, when an image is read fractionally quicker or a different diagnostic procedure is chosen at some point in the future, reimbursement means healthcare providers can capitalise on the availability and use of AI tools

More significantly, by facilitating the direct financial reward for use of AI tools for the likes of Optellum, as well as other vendors and solutions which have also secured reimbursement such as the stroke tools from the likes of Viz.ai, RapidAI and Aidoc for example, and the ultrasound guidance tools from Caption Health, the CMS is helping AI on its march to mainstream adoption. Such AI tools, which deliver greater value to radiologists, and in turn are rewarded by the CMS (in the US market at least), demonstrate a very powerful sales proposition and facilitate their use at ever greater numbers of sites.

This also heralds a new model for assessing market leaders. Instead of indirect methods for assessing a company’s place, such as its ability to fund raise or its ability to secure FDA approval, reimbursement is a more direct metric, showing clearly which solutions are being deployed and used clinically. By this metric, Optellum is punching well above its weight, with its funding deficit not necessarily a barrier to its rise to AI’s top table.

About Signify Premium Insights

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here