• Prostate cancer notes: ICD-O-3 coding

    I’ve written before that I use this blog as my research notebook. Trust me, it’s an extremely valuable way to preserve my (terrible) memory of details about what I’ve read. To the extent that readers who are also subject-matter experts chime in, it can also enhance the quality of my work. If you care about the subject, this is of benefit to you (and your professional community) too because it improves my treatment of it in future posts. In case you want to play along, I’m indexing questions in each post with the notation “[Q#]” where # is an integer. These are end-notes. Scroll down and see if you can answer any of my questions, if you wish.

    With that as general explanation, I’m going to be blogging some notes about prostate cancer. Some of these posts will be way, way down in the weeds and may bore the stuffing out of many readers. If you’re one of them, just move on. Note, however, that I will try to make the actual language of weedy posts humorous, because I use humor to distract from my lack of wisdom. Other posts may be at a higher level, more directly pertaining to policy or coverage decisions, so they’ll be more inherently interesting.

    This one is weedy. You’ve been warned.


    I’m beginning to get up to speed on prostate cancer morphology and data coding thereof. It appears to me as if the data standards are codified in ICD-O-3 documentation. (ICD = “international classification of diseases,” O = “oncology,” and 3 = third edition. And, yes, there is a Wikipedia page for it.) In particular, I’ve been “enjoying” the prostate cancer section of the ICD-O-3 SEER Site/Histology Validation List (pdf). (That’s ICD-O-3 topographical type C619 for those of you scoring at home.) A portion of that section looks like this:

    I will confess, I barely don’t understand cancer terminology yet, so I plucked out part of the “squamous cell” chunk of codes because “squamous” has an interesting sound, though probably not to those with squamous cell cancer, whose likely challenging relationship with the word I fully acknowledge. It makes me think of someone who is squeamish about being famous. “Oh he doesn’t talk to the press. He’s squamous.” Vastly more about squamous cell cancer here, only a bit of which I’ve read as of this writing.

    It’s pretty clear from the organization of the ICD-O-3 that you’ve got your topographical code (prostate = C619) and then you’ve got your “higher level” morphology code [Q1] (807 = squamous cell carcinoma). See how all the entries in the exhibit above start with “807”? But then there are two more numbers, one preceding and one following a forward slash (/). These indicate “lower level” morphology (or sub-types) [Q2]. The number immediately before the slash is an integer in sequence (0, 1, 2, 3, …). For prostate cancer, the number following the slash is always a 2 or a 3. Often, but not always, a 2 is associated with an “in situ” type, meaning, I think, that the cancer has not invaded surrounding tissue [Q3]. There are all manner words affiliated with sub-types: small, giant, clear, noninvasive, keratinizing, NOS, and some others. I don’t know what they all mean. I don’t know what makes a “/2” a “/2” and what makes a “/3” a “/3” [Q4]. I don’t know what work the slash does, though it suggests yet another level of typology (finer level after the slash) [Q5].

    Switching gears, another bit of information of interest in my research is a prostate cancer’s Gleason score. Trouble is, some of the data I will be using doesn’t have Gleason score coded. So, I’m wondering whether I can infer it (or, if not it exactly, at least to within some range) from other information. With that in mind, I noticed the following in the 2010 edition of the Facility Oncology Registry Data Standards (FORDS) manual (pdf). (There is a 2012 FORDS manual, as well as older editions.)

    Perhaps a range of Gleason score can be inferred from terminology and/or histolic grade, if the associations implied by the chart are standards that are widely followed.

    I’ll stop here for now and get back to my reading.

    Questions (see text)

    Q1: “Higher level” is probably not the right terminology here. What is?

    Q2: Again, what’s the proper terminology for these lower level codes?

    Q3: Right?

    Q4: Help?

    Q5: Yes?


    Comments closed
    • From a pathologist:

      Hi – there is essentially no such thing as a Gleason score 2-4 prostate cancer. This is in part a vestige of the past, Gleason’s pattern 1 no longer really exists as a malignancy. Pattern 2 exists but is not that common as an isolated lesion.

      Anything on biopsy nearly automatically receives at least a pattern 3 (thus, 3+3=6). Very rare occasions can lead someone to give something a Gleason 3+2=5 on biopsy. You can of course see something with Gleason pattern 2 on resection but that is almost never the only tumor. Even resections with Gleason score <6 are extremely uncommon. The Gleason system is so entrenched now that there is little likelihood that the 1-5 system will be compressed into a new 3-tier system with patterns 1-2 removed. Thus, 99% of cancers reported are Gleason score 6-10.

      In a sense, the clinical cut points are
      1) Gleason 6 or less is well differentiated, very unlikely to kill the patient
      2) Gleason 7 is potentially more aggressive and typically needs treatment, although even here there are some who argue there is a major difference between 3+4 and 4+3.
      3) Gleason 8-10 is aggressive and potentially fatal if untreated.

      Q3 is correct – in situ has no invasion. There are some pathologists who will report things in the cervix, breast, lung, or even bladder as "in situ cancer with focal microinvasion" which basically means it is not in situ anymore, but barely. What the means clinically who knows.

      Prostate cancer though has HGPIN which is not really the equivalent of carcinoma in situ – more like an atypical hyperplasia.

      I don't know what the /2 and /3 mean either. But it looks to me as though "3" refers to an invasive carcinoma, and 2 refers to in situ or non-invasive carcinoma. In-situ carcinoma means different things in different organ systems. An in situ breast cancer can still be aggressive and require surgery + radiation, whereas in situ cancer of the colon, for example (in a polyp) requires nothing but removal and surveillance.

      It looks to me like "/0" refers to benign tumors (like a lipoma) and "/1" refers to tumors that are probably benign but on occasion can become malignant or become locally aggressive – like uterine smooth muscle tumors or certain types of meningioma. Not technically "cancer" but potentially problematic.

      The range of stuff included in "/3" is extreme – from low grade malignancies like Gleason 6 prostate cancer to glioblastomas and metastatic pancreatic cancer.

      • Thank you!

        Though I didn’t mention it in the post, I’m going to be looking at data from 2000-2004. Would I see reference to sub-6 Gleason scores in that era?

    • Possibly. In most of the literature I see, prostate cancers are either grouped with a category of “Gleason score less than 6” or “Gleason score 6 or less.”

      However, I tend to be very skeptical of any study that includes a separate category of “gleason score less than 6” because that is so rarely diagnosed these days (at least, by competent pathologists). That trend been going on for at least 10 years now. I can’t really claim expertise on what percentage of cancers would fall under the “Gleason less than 6” more than 10 years ago, but these days it is less than 1% of cancers. Gleason 6 can account for a sizeable portion though.

      Compounding the problem is that Gleason grading has shifted over the years. Some of the patterns of cancer growth that were previously called “Gleason pattern 3” have been re-classified as pattern 4 because they have been recognized as more aggressive. For an example, see here: http://www.aqch.com/feature/2008/feature022008.php

      The major revision occurred in 2005 officially, but a lot of this movement was happening before then particularly among expert/experienced urologic pathologists.

    • Per the Wikipedia page, “NOS” is “not otherwise specified”. This indicates a “classification of last resort”, to be used only if no other classifications match. The other words in the descriptions (besides NOS) are histological jargon.

      Q4 and Q5. From the WHO page (linked in the Wiki page):-

      “The first four digits indicate the specific histological term. The fifth digit after the slash (/) is the behaviour code, which indicates whether a tumour is malignant, benign, in situ, or uncertain (whether benign or malignant).”