Farewell Pluto – Farewell LCIS

Poor LCIS. Downgraded. Belittled. Banished to revolve around the internet forever, with no respect and no identity. Not even a proper name. No more Stage 0. At the stroke of midnight Dec. 31, 2017, as the ball drops in Times Square, we will drop the ball on LCIS with the 8th edition of the AJCC Staging System. LCIS will cease to be “cancer,” and will emerge on New Year’s Day as…hmm…LCIS? Lobular carcinoma in situ? A benign lesion still called “carcinoma?”

If there’s a newly proposed name change, I missed it. “Lobular neoplasia” was tried in the past, but didn’t catch on, maybe because it dragged with it the oft-indistinguishable ALH. Should we revisit “neoplasia?” After all, Dr. David Page’s ALH is often Dr. Paul Peter Rosen’s LCIS. Those darned continuums!  (in photo above, clear-cut LCIS is at the top, but in the middle of the slide, if this were the only abnormality, some would call it ALH while others LCIS.  A normal lobule is at the bottom of the slide).

Since Foote and Stewart’s coinage of the term “LCIS” in 1941, the lesion has enjoyed the same respect as DCIS, both considered to be non-invasive cancers. But no more. The two are parting ways, a division that the NSABP stamped with approval in the 1990s when the P-01 trial for tamoxifen prevention included LCIS patients as “high risk,” while DCIS patients were shuttled elsewhere into the DCIS trials.

But is there really that much difference between LCIS and low/moderate grade DCIS? Both have approximately the same future risk for invasive breast cancer, with perhaps the only difference being the unpredictable location of that event for LCIS patients. But in terms of threat to a patient’s life? Not much difference.

Poor Pluto was dragged into planetary oblivion when scientists realized the untenable position that Pluto didn’t fit the definition of a planet. Sure, it was spherical in shape by virtue of its own gravitational forces (old definition), but it didn’t “dominate its neighborhood” (new definition) whereby a planet sweeps up smaller bodies in its orbit (not to mention that Charon, one of Pluto’s moons, is so large in relation to its planet that the neighborhood had two kingpins, plus several henchmen). Size was problematic already, given that other bodies in the solar system are close to Pluto’s volume, but the non-planet insult was not based on size alone. For astronomers, the solution was simple – a change in taxonomy. From planet to dwarf planet. It turns out, of course, that the number of dwarf planets may be, well, astronomical.

Under the new system, a 1.0cm low grade DCIS will still be called Stage 0 “cancer,” but a 5.0cm area of pleomorphic LCIS, far more threatening to the patient, will be nothing more than a pitiful Pluto trying to sweep clean its orbit in order to get reinstated. The breast cancer expert panel that took the most revolutionary steps ever in the new 8th edition of AJCC staging gave a nod to pleomorphic LCIS as a distinct entity in the text of their opinion, but one gets the impression that the broad array of LCIS presentations was simply too much to handle, especially when these more ominous sub-types of LCIS are relatively rare.

It may be an oversimplification to designate all non-classical LCIS as “pleomorphic.” Of course, many sites on the internet don’t even bother with this basic distinction, calling all LCIS a mere “tissue risk” with a 15% chance for future breast cancer (true, if we’re only talking about the next 10-15 years). But I think the pendulum swung too far, largely because so many cases of LCIS would be called ALH on review. With ALH, the “tissue risk” moniker applies, with implications indistinguishable from ADH when considering long-term estimates for developing invasive breast cancer.

Many years ago, Dr. David Page attempted to define what made certain instances of LCIS more of a threat for invasion, and his most consistent finding was in the qualitative aspect (not so much in the amount of LCIS). That is, when acinar units are so bloated with cells that they appear to be on the verge of exploding (my description), the risk of future invasive cancer is higher, even if the finding is only present on a single slide, single focus.

Indeed, modern reviews of natural history, such as the recent landmark effort recently by Tari King, et al ( J Clin Oncol 2015; 33:3945-3952) are more restrictive to clear-cut cases of LCIS. In Dr. King’s analysis of 1,004 patients opting for surveillance after a diagnosis of LCIS at Memorial Sloan-Kettering, the calculated future risk (including development of DCIS ) was a 2% annual incidence, much higher than the 1% per year commonly quoted. In addition, the future cancer occurred on the same side as the LCIS in 63%, plus an additional 12% were bilateral. Invasive lobular cancer was the culprit in 27% of patients, disproportionate to the 5-10% of lobulars in the general population. All of this points to non-obligated precursor behavior, not simple “tissue risk” as one sees with ALH, ADH and, to a lesser extent, other proliferative pathologies.

So what’s really going on under the microscope that is failing to make it to the clinic? Our colleagues in pathology have been trying to tell us for a long time that there are a wide variety of findings that one can see with LCIS — that is, LCIS is not a single entity. Various classification schemes have been proposed, such as:

 

Classic LCIS, Type A – low grade nuclei, inconspicuous nucleoli

Classic LCIS, Type B – larger nuclei, small nucleoli

Florid LCIS (a.k.a LCIS with central necrosis/calcifications) – massively expanded acinar units, but this designation does not require high grade nuclei, and may be a distinct clinical picture from pleomorphic LCIS. This type of LCIS would have been called DCIS in the years prior to the introduction of e-cadherin.

Pleomorphic LCIS – high grade nuclei and prominent nucleoli (central necrosis not an integral feature)

Apocrine Pleomorphic LCIS – as above, but with eosinophilic granules in the cytoplasm, and other features of interest only to pathologists.

(And this doesn’t even take into account questions about the significance of other morphology issues such as clear cells, signet ring cells, etc. nor the Pagetoid spread beneath ductal epithelium that can occur while still in the ALH phase.)

 

Are there clinical implications? It appears so. Problem is, the “non-classic” lesions are infrequent and keep us from amassing solid data. The current impression is that “florid” LCIS is associated with a worrisome rate of synchronous (and metachronous?) invasion, while the primary concern for pleomorphic LCIS is with regard to the future development of a high-grade pleomorphic invasive lobular lesion, a “lobular” which doesn’t play by the rules (even though the risk of conversion to invasive disease may not differ substantially from classic LCIS).  Perhaps with more data, we can lump the “non-classic” LCIS lesions into a single entity. But for now, the “lumpers” need to let the “splitters” iron things out.

The many faces of LCIS are analogous to the various parameters that prompt the multiple classification systems proposed for DCIS – the difference is the rarity of “florid” and “pleomorphic” types of LCIS. If these lesions were as common as high-grade DCIS, we wouldn’t be witnessing the current demotion of LCIS unless we banished all forms of DCIS along with it (which we might see in the AJCC manual, 9th edition).

The “tissue risk” moniker for LCIS was based on that era (late 1980s and early 1990s) when the observational studies matured, indicating that the future cancer was likely to be on either side with equal frequency, completely unrelated to the old biopsy site, and that it would be ductal, far more likely than invasive lobular – thus, not a direct precursor.

But times have changed. Evidence continues to accumulate that the ipsilateral side is more likely to develop invasive cancer than the contralateral side, and that there is a disproportionate number of invasive lobular cancers, sometimes occurring at the old LCIS site. This is not “tissue risk” behavior, but is “non-obligated precursor” behavior, much like low grade DCIS. (From personal experience, once you perform a wide excision for LCIS, then 10 years later, you’re back in the O.R. again, operating on invasive lobular carcinoma originating at the exact site where you previously did the biopsy, you quit using the phrase, “LCIS is only a risk factor.”)

Molecular studies also support the precursor notion, with a lineage of mutations that can be tracked from LCIS to ductal cancer with as much facility as LCIS to invasive lobular cancer. Indeed, these molecular findings have been reflected in the clinic for years, but we tend to put blinders on. Recall how often invasive ductal cancer comes with a background of LCIS instead of DCIS (and we shrug our shoulders at conference and move on to the next patient). Or, as we saw twice in one week here at Mercy-OKC while I was writing this editorial – 2 distinct invasive cancers in close proximity, one ductal, one lobular, but both surrounded by a background of LCIS.

So what is LCIS? And what should we call it?

I don’t have the answer. I only have questions. “Needs more data” has been our cop-out for many decades now. But from a philosophical standpoint, maybe the problem is not in our lack of understanding or even our lack of data (we generally know what LCIS portends for the patient). Our problem is semantics. We simply don’t have a word or phrase that captures what we’re trying to say.   Noninvasive cancer or in situ carcinoma has been considered too harsh because the C-word is still there. On the other hand, Dr. Esserman’s catch-all term, “indolent lesion of epithelial origin (IDLE),” sounds like the problem is so low on the worry chart that it hovers in the same zone as ordinary hyperplasia. Of course, we’re all hoping that molecular studies give us a new classification system some day and a new terminology that speaks to everyone.

In the meantime, how about a descriptive name focused on behavior rather than morphology? The only one that fits for LCIS is: “Non-obligated precursor with a tendency toward a bilateral geographical distribution. Or, NOPTTBGD.

Ugh. I give up. The problem is just as real for DCIS as it is for LCIS. Let’s just call them all dwarf cancers – and be done with it.