When considering purchase of a Photon Counting CT for postmortem imaging [EU MDR]

The EU CE marking regulation for medical devices, including CT scanners, is primarily governed by the Medical Devices Regulation (EU) 2017/745 (MDR). This regulation outlines legal requirements for the safety and performance of medical devices placed on the EU market.

A number of  issues appear to exist with Photon Counting CT scanners for forensic / medicolegal use.

• Due to build/hardware aspects, Photon Counting CT technology seemingly tends to produce relatively high noise data, whose quality still may exceed current clinical CT scan quality at low CT doses as used clinically; due to post mortem CT use related absent CT dose restrictions, however, a Photon Counting CT currently may appear to constitute a disadvantage for forensic / post mortem / deceased examinations with that regard, particularly when used in a time critical manner. Practical relevance of its potentially superior data/image quality delivery at CT doses above 700 or 1100 mAs with a similar speed as a current CT scanner with energy integrating detector delivers is yet to be proven.
• Anecdotal reports stated that a useful whole body post mortem CT would take over 3 hours to reconstruct, and another anecdotal statement from developer sources said it was unlikely these reconstruction times would become less over the next years. Reconstructing noise reduced data using various levels of CT energies, over hours and days, certainly may have a certain degree of research appeal, but who else already spent hours averaging loads of data for less noise may not find that exceedingly fascinating.
• The noisy (or smoothed denoised) Photon Counting CT data may be supplemented with very high pixel/voxel resolution, whereas the feature, the finding, the focal pathology or trauma finding itself, may not necessarily experience a similar resolution improvement; if the user is lead to believe to deal with high resolution data simply  because of higher pixel/voxel count and larger data size, then we may not be dealing so much with physics of high feature resolution, but neurophysiology of consumer device use; a current CT’s resolution of an isotropic 0,6mm voxel size blown up to an isotropic 0,2mm voxel size will correlate with a 3 x 3 x 3 multiplication of data size, so you may wait 27 times as long for, say, data transfer. 
• The reconstruction computations across various different initially noisy scans, possibly available in different CT energies, may reduce noise alone also because all combination of different scans tends to reduce noise, but a single whole body post mortem CT may take a few hours of reconstruction time, easily.
• So in short, data quality and scan/reconstruction time seem to be relevant issues, if not road blocks.
• Photon Counting CT is both an idea, an ideal, and a promise, and vice versa these imaginations, an attempt at implementing this as a particular instance. These are different, in that the ideal is great, superb, futuristic, but the actual implementation instance may even feel like a nightmare, given specifics, of course.
• This is not specific to any new technology, in fact, it pertains to typical user experiences: depending on specifics, a new smartphone may feel like a nightmare compared to an old mobile phone that had its battery run for days without recharging; an electric car may be a nightmare to drive given recharging options, compared to a car with a good old diesel motor, and so on. Neither is the idea, the ideal, bad, nor the plan to consider a new technology – however, one should be mindful about this. In cases where the manufacturer is obliged to declare specifics, we expect any such information to be presented automatically.
• “The spectral performance of PCCT in material-specific imaging (eg, iodine imaging, VMI) may not necessarily be superior to that of conventional CT–based dual-energy imaging at the same radiation dose (95), mostly attributable to spectral distortions caused by several nonideal properties of the photon-counting detector, such as pulse pileup and charge sharing (96,97).” [link] [1]
• Detector materials play an enormous role in what can be implemented. There, Cadmium Tellurid is a material whose crystal growth seems to be fraught with difficulties to obtain a pure material. However, purity is required for good performance. Also, CdTe is reported to be highly toxic. Conversely, Silicone can be obtained with great purity and isn’t nearly as toxic. The race for best material as implementation for the ideal of Photon Counting CT promises is on.
• To add on to that, when we expect new detector materials, one big question is whether the detector can be swapped: so you buy a CT scanner, and the question is, if they build a new and better detector: can the detector be swapped?  Or do we always just buy a new gantry? May the detector material toxicity play into the detector swap being an option at all? And how – from view of curiosity, and, from a medical device regulation view — can these obvious technology inherent realities be left out from promotional materials?

Any user or customer, the clinician or radiologist, is not the designated person to professionally point out any of that. That is not because we are generally unable to speak, but, because medical device law requires the manufacturers to also and without prior indication, by themselves, not from own laudable initiative but from a pure legal compliance standpoint. therefore, declare the potential risks, use limitations, issues, themselves, also and in particular in their promotional materials, without the users having to ask. As side remark I may add that when I ask a manufacturer “please send images to demonstrate the allegedly superior soft tissue contrast of your new device”, which if the device performs as advertised they should have a ton of, and the manufacturer representative then responds with “I am not of your opinion” to that exact question, but not includes any images, like, at all, then we are definitely at an awkward spot not only with regard to human communication, but also with regard to their MDR compliance.

Here are key points regarding the regulation and how they apply to CT scanners and similar medical devices:

1. Compliance and Safety

The manufacturer must ensure that the CT scanner meets the general safety and performance requirements outlined in Annex I of the MDR. This includes ensuring the device is designed and manufactured to ensure a high level of protection for users and patients.

2. Documentation and Declaration

The manufacturer is required to prepare comprehensive technical documentation that demonstrates the device complies with the MDR requirements. This documentation includes:

• Risk assessment and management processes.
• A detailed description of the device and its components.
• Clinical data to support the device’s performance and safety.
• Labeling and instructions for use, including any potential risks or limitations.
• The Declaration of Conformity, stating the device complies with all relevant EU regulations.

3. Post-Market Surveillance

Manufacturers must also implement and maintain a post-market surveillance system to continually monitor the performance and safety of the CT scanner once it’s on the market. This system is used to report any incidents or issues, allowing manufacturers to address potential risks in real time.

4. Issues, Problems, and Limitations

Yes, the manufacturer is obligated to declare, also with promotional materials, any known issues, problems, or limitations associated with the CT scanner. This includes:

• Potential risks: The technical documentation and promotional materials must provide clear warnings about risks related to the use of the CT scanner, such as radiation exposure, calibration requirements, or other safety considerations.
• Limitations in use: If the device has specific limitations (e.g., certain patient populations or contraindications), these must be disclosed in the documentation and promotional materials.
• Adverse effects: Any known adverse effects or situations where the device may not perform optimally must be disclosed.

5. Promotional Materials

Under the MDR, promotional materials must not mislead the users regarding the device’s intended purpose, safety, or performance. Specifically, Article 7 of the MDR prohibits the following ¹:

• False or misleading claims about the device’s safety and effectiveness.
• Overstating the benefits of the device.
• Failing to disclose known risks, issues, or limitations.

6. Instructions for Use

The instructions for use (IFU) must be clear, complete, and include all information necessary for the safe use of the device. This includes details on:

• Intended purpose of the CT scanner.
• Clear instructions for installation, maintenance, and calibration.
• Warnings and precautions, including any known device limitations or operational challenges.
• Information on how to report issues and how the user can participate in post-market surveillance efforts.

Summary

Manufacturers of medical devices like CT scanners are required to provide comprehensive documentation and promotional materials that transparently disclose any issues, problems, or limitations.

The goal of the EU MDR is to ensure that users are fully informed about the safety, effectiveness, and limitations of medical devices to mitigate risks and protect patients. This includes clear labeling, accurate promotional materials, and a robust system for post-market surveillance.

We are looking forward to developments of Siemens, Naeotom Alpha, as well as General Electric / GE Photon Counting CT, and other manufacturers. We expect their presentations, promotional materials as well as documentation to be CE conformant, however, therefore to explicitly also point out limitations and issues. If you read that here first, though, then, hm.

@article{fletcher2024photon,
  title={Photon-counting CT in thoracic imaging: early clinical evidence and incorporation into clinical practice},
  author={Fletcher, Joel G and Inoue, Akitoshi and Bratt, Alex and Horst, Kelly K and Koo, Chi Wan and Rajiah, Prabhakar Shantha and Baffour, Francis I and Ko, Jane P and Remy-Jardin, Martine and McCollough, Cynthia H and others},
  journal={Radiology},
  volume={310},
  number={3},
  pages={e231986},
  year={2024},
  publisher={Radiological Society of North America}
}