"Index tumor" vs "Index lesion"

[Andrey Fedorov]

Hi,

I have a question about "Index tumor" term http://www.radlex.org/RID/RID11517.

Is it appropriate to use this term to define the index prostate lesion, per PI-RADS guidelines?

Note that PI-RADS assessment specifically calls for evaluation of "index lesion", not "index tumor" (i.e., see http://radlex.org/RID/RID50295 "PI-RADS Lesion Assessment Category").

Thank you

Andrey Fedorov

[Ken Wang]

Hi Andrey,

Thanks for your message, and your question.  The situation you are identifying is a bit tricky.  I would say that "index tumor" is probably not a great direct synonym for a PI-RADS "index lesion".  On the other hand, while we do have many terms from PI-RADS, I note that "index lesion" is not specifically listed in the PI-RADS lexicon document, and we often rely on such documents to determine what terms to mark as formal terms from a given source.  However, I do recognize that the full PI-RADS document does utilize the term "index lesion" as you say. 

All of that said, I'm copying additional members of the RadLex team here, and we will investigate.  (Ross and Beverly, perhaps we can add an agenda item for the next meeting to discuss PI-RADS?)

Thanks,

Ken Wang

--
You received this message because you are subscribed to the Google Groups "RadLex Feedback" group.
To unsubscribe from this group and stop receiving emails from it, send an email to radlex-feedba...@lists.rsna.org.
To view this discussion on the web visit https://groups.google.com/a/lists.rsna.org/d/msgid/radlex-feedback/2812edb2-3961-42e8-87ba-0d0f9a0cfce8n%40lists.rsna.org.

[Andrey Fedorov]

Ken,

Thank you for the prompt response! 

To give you some the context, this came up in the process of developing Structured Reporting capabilities in the DICOM standard to support prostate MRI reporting (it is a very large and unwieldy document, with many unresolved edits and comments, but it is in the public comment stage, and just in case anyone is interested to see it - you are welcome: https://docs.google.com/document/d/1C9zQF1wmXasVrM73OhBLQNb5MqWT2l16-nVGcIXhXZs/edit?usp=sharing).

The process of developing DICOM reporting templates involves codification of concepts and values, and RadLex has been essential in providing many of the terms from PI-RADS. I initially missed RadLex "index tumor" completely, and instead proposed a new code to define "index lesion" (which would be a new DICOM code, if/when the supplement becomes part of the standard), but someone did notice "index tumor", which triggered my question to you.

I am looking forward to get further guidance on this question!

Andrey

[Andrey Fedorov]

Ken,

if you meet to discuss PI-RADS "index lesion", then here's another two terms in the context of PI-RADS.

PI-RADS lexicon includes the term "encapsulated", which is defined as "Bounded by a distinct, uniform, smooth low-signal line (BPH nodule);  completely  encapsulated  nodule  is  entirely  surrounded  by  a  smooth  low-signal  line  in  at  least  two  imaging  planes  (“typical  nodule”);  almost  completely  or  incompletely  encapsulated  nodule  is  not  entirely  surrounded  by  a  smooth  low-signal  line  (“atypical  nodule”)", which indirectly defines "typical" and "atypical" types of nodules in that specific context based on how the nodule is surrounded by a "smooth low-signal line".

While developing the DICOM PI-RADS capabilities, we received a specific suggestion phrased as follows: "Additional qualifiers from PI-RADS 2.1 such as mostly encapsulated are currently missing".

Do you have any comments on how RadLex could be amended to include those additional qualifiers ("completely encapsulated" vs "incompletely encapsulated")?

Thank you

AF

Boundedbyadistinct,uniform,smoothlow-signalline(BPHnodule); completely encapsulated nodule is entirely surrounded by a smooth low-signal line in at least two imaging planes (“typical nodule”); almost completely or incompletely encapsulated nodule is not entirely surrounded by a smooth low-signal line (“atypical nodule”)

[Ken Wang]

Hi Andrey,

Thanks for your additional questions.  As Ross mentioned, we will discuss at our next meeting.  I appreciate the questions you raise, because these are the kinds of challenges that we struggle with and which we try to address.

As for "encapsulated", "completely encapsulated", "incompletely encapsulated" or even "almost completely encapsulated" or "mostly encapsulated"... these could all be assembled into a micro-ontology under RID49498 "encapsulated margin".  That is, we could say that "encapsulated" is a synonym for "encapsulated margin", and then create "completely encapsulated" as a child under "encapsulated margin" in the subsumption hierarchy, etc.  But we would want to be able to distinguish these variants by way of definitions, if possible.  Do you know if PI-RADS gives definitions for what constitutes "complete" vs. "incomplete"?  The text of your message refers also to "almost completely" and "mostly"... are these also distinct formal PI-RADS terms?

An alternative approach is to use post-coordination, whereby multiple RadLex terms could be assembled and referenced by DICOM to create more sophisticated expressions.  For example, RID5686 "incomplete" could be used as part of this approach.  However, that term is of course quite generic, and the specific meaning in a given usage would depend on the context of the post-coordination rather than on a specific definition maintained in RadLex in that case.

Anyway, feel free to let me know if you have any thoughts about the above, and thanks again for getting in touch.

-Ken

[Andrey Fedorov]

Ken,

I do not think PI-RADS defines what constitutes "complete" vs
"incomplete". What I included in my email was verbatim text from the
(latest) PI-RADS 2.1 document lexicon (p.33 of the PDF linked from
https://www.acr.org/Clinical-Resources/Reporting-and-Data-Systems/PI-RADS).

I copy Dr. Macura, who (at least in PI-RADS 2.0) was leading the
lexicon related efforts. It would be great to have her thoughts on
this.

Andrey