Quality and Outcomes Research in Head and Neck Cancer: From Bench to Bedside with Yale Leading the Charge

October 19, 2022

Information

Yale Cancer Center Grand Rounds | October 18, 2022

Presentation by: Dr. Saral Mehra

ID8178

To CiteDCA Citation Guide

00:00Hi. Welcome everybody.
00:01I'm Barbara burtness. I'm a.
00:03Medical oncologist and work on
00:05head and neck cancer and I'm
00:07really like I could not be more
00:10extraordinarily delighted than I
00:12am to be today presenting Dr Saral
00:15Mera as our grand round speaker.
00:17Doctor Mehra is an associate professor
00:19of surgery in otolaryngology and
00:21Section chief of head and neck surgery.
00:23He received his medical degree from
00:25Columbia University College of
00:27Physicians and Surgeons in New York,
00:29also obtained an MBA there and went
00:31on to residency in Otolaryngology
00:33Head neck surgery at New York
00:35Presbyterian Memorial Sloan Kettering,
00:37and then completed his training with
00:40a extremely coveted fellowship in
00:42head and neck and thyroid cancer
00:44surgery at Mount Sinai,
00:46including Subspecialization in
00:48complex reconstruction.
00:50His clinical practice focuses on
00:52treating patients with head and
00:54neck salivary and thyroid diseases,
00:56particularly those patients who
00:58need a very advanced resection
00:59or advanced reconstruction.
01:01And I've been privileged to share
01:03many hundreds of patients with him.
01:04And I can like very personally say
01:07his oncologic and reconstructive
01:09outcomes are exceptional.
01:11And so his primary research
01:12interest ties in this very nicely
01:14because he focuses on measuring and
01:16improving quality in the treatment
01:17of head neck cancers and I think
01:19one of the central conundrum.
01:21Our field is how do we take exceptional
01:24care and and broaden its accessibility
01:27and broaden its applicability.
01:28So he's going to be talking about
01:30quality and outcomes research and
01:32head neck cancer from bench to
01:34bedside with Yale leading the charge,
01:35Cyril.
01:41Thank you so much for that
01:43introduction Barbara and thank
01:44you everybody for logging in and
01:46and and listening to this talk.
01:48I know it's a diverse group of basic
01:51scientist clinical researchers
01:52and hopefully some outcomes
01:54researchers out there as well.
01:56So that's my topic.
01:59It's been discussed.
02:00Just make sure this is working.
02:14OK, so I have no I have
02:16no financial disclosures,
02:17so I couldn't do a head and neck
02:19research talk without mentioning 2
02:22important studies recently published,
02:24ECOG 3311 and keynote O48,
02:25of which Yale specifically Doctor
02:27Burtness was lead author on both of
02:29these very different types of studies.
02:31But these are practice changing
02:33studies recently published.
02:35Um ECOG 3311 looked at T1 and T2
02:38resectable HPV positive oropharynx
02:41squamous cell carcinoma and
02:44keynote O48 looked at locally
02:45the opposite end of the spectrum,
02:48locally incurable recurrent or metastatic
02:50head neck squamous cell carcinoma.
02:52What I've put up here,
02:54I'm not going into details of these studies,
02:56but what I've put up here are
02:59survival curves from these huge.
03:01Hugely important recent practice
03:03changing studies, OK,
03:04I'm not going to get into details
03:07about these over the next 45 minutes.
03:08What I really want to do though is
03:10convince you of two things and I think
03:12I may have already done the first one
03:14because it doesn't take much convincing.
03:16Large scale multi institutional
03:18randomized control trials are important
03:20and they can lead to practice changing
03:23innovations in the care of cancer patients.
03:26Any cancer, I'm going to put a
03:29check mark beside that already.
03:31Based on that first slide and
03:32which you probably already know,
03:33but really what I want to do
03:35is tell you that the quality,
03:37that quality of care research
03:39or call it what you will,
03:41patient reported outcomes reaches
03:44patient centered outcomes research.
03:47Comparative effectiveness research goes
03:48by different names at different times.
03:51But when you can take this from
03:53the bench to the bedside,
03:55we can have a remarkable impact
03:57on patient outcomes.
03:58And not only that but this type
04:00of research is actually necessary.
04:02For I'll put in quotes the real
04:04science to mean anything.
04:06That's the basic scientist,
04:08the clinical trials.
04:10That's what I want to convince you of.
04:11By the end of this talk,
04:13I'm going to start with a case.
04:15This is a 52 year old man.
04:16This is actually very recent.
04:20And within our system,
04:21OK, the very recent case,
04:2352 year old man never smoker 2.9
04:26centimeter right neck mass and a
04:28right tonsil mass was seen in our
04:31system and had a right tonsil mass.
04:33Biopsy showed poorly differentiated
04:35squamous cell carcinoma and the P-16
04:38stain was strongly and diffusely positive.
04:41The patient went for surgery.
04:45And what the pathology identified with
04:47the tonsillectomy and a neck mass
04:49excision was a P-16 positive cancer,
04:523 centimeters extending to the
04:54margins of resection and a 2.7
04:57centimeter node with no ENE.
05:01And a few other nodes also that
05:03that came with the specimen.
05:04So this patient was then referred
05:06for radiation and chemotherapy.
05:07There's a positive margin.
05:10The guidelines say radiation
05:12chemotherapy was then re referred
05:15to but by care center physician
05:18to the head neck disease team at
05:20the at at the Big House.
05:22And this was interesting because
05:24a few things first of all was
05:26staging completed prior to treatment
05:27was one of
05:28the questions that came up
05:30when they were discussed.
05:31And the answer was no.
05:32And then more important was the
05:35latest science applied to this patient
05:38ECOG 3311 had already come out.
05:39I told you this is a very
05:42recent case and preoperatively
05:43pre surgery this patient will
05:44be in the low risk category of
05:48T1N1 cancer where we
05:52could potentially operate.
05:54And go to observation, no radiation,
05:56no chemotherapy and this patient
05:58have a 94% progression free survival
06:00based on the data from ECOG 3311.
06:03But now this patient has a positive margin.
06:06Now he's in the high risk category
06:08and now he technically should go
06:11on to chemotherapy and radiation.
06:13Nothing about this tumor change,
06:15nothing about the patient changed,
06:16about the biology changed,
06:18but now the treatment recommendation
06:21could be different.
06:22So we had a long discussion in tumor board.
06:25And what we decided to do after
06:27multidisciplinary team care was complete,
06:29the staging taken for transoral
06:32robotic surgery to clear that margin
06:34and do a formal neck dissection.
06:37Unfortunately,
06:37this patient did have to escalate
06:39after a discussion,
06:40we did decide to escalate his therapy
06:43to radiation because of the original
06:45positive margin and discussing the
06:47risks and benefits with the patient.
06:50Controversial,
06:50but that's what we decided to do.
06:53OK. So that's my case.
06:55Some kind of frames,
06:56this discussion,
06:57so some background and definitions.
06:59What is quality in cancer care?
07:01I'm not going to read all these,
07:03but if you look at some of the literature
07:06on quality and cancer care nationally,
07:09they have all these fancy definitions.
07:11But this is what I think is just,
07:13it's so simple,
07:13so simple and you all know this already,
07:15I know, but it's just getting the
07:17right care to the right patient at
07:20the right time and doing it every time.
07:22That's it, that's quality.
07:24In healthcare,
07:25especially cancer care.
07:26So what went wrong with this patient?
07:29This patient is going to do fine.
07:30The survival rate still 90% plus,
07:33you know,
07:33it wasn't technically harmed other
07:35than treatment and functionally
07:37and maybe some side effects,
07:39but he's going to live,
07:40he's going to have a good outcome.
07:42But,
07:42but I would argue this wasn't the
07:45highest quality of care and it
07:47wasn't applying the latest science.
07:49So if anyone here has been
07:51to Business School,
07:53you would have read a lot
07:55about Edward Edwards Deming,
07:57who created basically the Toyota,
08:00Toyota Quality Improvement
08:01process and brought you know,
08:04car production to the
08:05highest quality in Japan.
08:07And I think he said it best and
08:09it really applies to cancer care.
08:12What the fault in the interpretation
08:14of observations seen everywhere
08:16is just suppose that every event,
08:18any defective mistake and accident
08:20is attributable to someone,
08:22usually whoever's closest at hand or
08:25some special extraordinary event.
08:28But the fact is,
08:28it's not the people that are the issues,
08:30it's the system that doesn't work
08:32when there's errors and mistakes.
08:34The Institute of Medicine,
08:36in their crossing the Quality Chasm report,
08:39said it well that poor quality care
08:41occurs not because of a failure of
08:44goodwill or knowledge or effort or resources,
08:47but because of fundamental shortcomings
08:48in the way cancer care is organized.
08:51There's no there's no bad doctor,
08:53there's no one who wants to harm a patient.
08:55That I know of at least.
08:57But the system can be the problem.
09:01So how do we measure quality?
09:02There's really three ways what a
09:04lot of us do, and what was done
09:07all the time, many years ago,
09:10was implicit or expert review.
09:13I'm in an ivory tower.
09:14I say it's high quality.
09:15I'm. I'm great.
09:16And so it is. That's it.
09:18But there are scientific
09:19ways to measure quality,
09:21specifically structure and process metrics.
09:26So why is structure so important?
09:29Take a moment to talk about this.
09:31So right now,
09:32I'm sure you can all recognize the
09:34organization of cancer care most of the US,
09:36the patient at least in the middle,
09:39and then doctors all around them
09:41and communication can be very,
09:42very challenging.
09:43This happens to be from a publication.
09:44And regarding thyroid cancer cares,
09:46we've got nuclear medicine specialists,
09:48we've got radiologists,
09:49we've got endocrinologist,
09:50medical oncologists and the patients.
09:54Left in the middle and can be quite confused.
09:58This is the structure of hospital
10:00care in the United States.
10:02There's about 6000 hospitals
10:04registered in the United States.
10:06About 1500 of them are registered by
10:09the ACS Commission on Cancer and about
10:1371% are treated at ACS Commission
10:16on cancer accredited hospitals.
10:20And I'm sure you've heard of SEER,
10:21which is a representative national database
10:24and that captures about 14% of cancer cases,
10:27but is supposed to be and probably
10:30is representative of cancer care
10:32within the country.
10:33So you can get incidence rates from Seer,
10:35where you can't get that from NCDB,
10:37which is a,
10:38which is the Commission on cancer data set.
10:42There's a number of
10:43comprehensive cancer centers,
10:44there's cancer centers,
10:46there's community cancer centers,
10:47all designated by the NCI.
10:49And I tried to find how many surgeons
10:51or medical oncologist radiation
10:53oncologists treat head and neck cancer.
10:55As you could guess,
10:57it's pretty much impossible to find that.
10:59So the organization of US cancer
11:02cases so diffuse diverse that
11:04providers have not you as a provider,
11:06if you are a doctor,
11:07have not had to provide accountability
11:10for specific processes or outcomes for
11:13you specifically or at your center.
11:16OK, so now let's get into some research,
11:18research and cancer quality
11:20that that's my background.
11:21So my research journey at Yale
11:25really is about quality of care,
11:28patient centers, out centered outcomes.
11:30And what I'm trying to do now is
11:33my version of bench to bedside.
11:35So how do we take the science in outcomes
11:38research and apply it to patients?
11:40It's really not just fluff.
11:41I know there's real quote real
11:44scientists out there, basic scientists.
11:46Huge randomized control types
11:49of signs is out there.
11:50I think this is real and going
11:52to try and convince you of that.
11:54So my foray into patient centered
11:56outcomes research started in fellowship
11:58and these republished in my first
12:00year being here some pretty large
12:02reviews on thyroid cancer care and
12:04how can we measure and improve thyroid
12:06cancer care across the country.
12:08And I reviewed all the databases
12:11for thyroid cancer and
12:12published these in thyroid with
12:14a nice team of colleagues.
12:16Um, kind of around the country
12:18and thyroid cancer care.
12:19When I got to Yale, I saw,
12:21Oh my God, there's actually a
12:22lot of people doing this stuff.
12:23People like Dan BofA and Ben Judson,
12:26who and and Kerry Gross,
12:28who were really nationally known
12:30in quality of cancer care research,
12:32two of them being surgeons through
12:34basic and head and neck surgeons.
12:37So what I wanted to do first was,
12:38well, how are we doing it, Yale?
12:40I mean, are we delivering quality
12:43care at Yale? I had no idea.
12:45Nobody had any idea. Yes.
12:46Many of these are even Dan and Ben.
12:48And you asked them,
12:49hey, how's your quality,
12:50how's your divisions quality,
12:51what's your margin rate,
12:53survival rates for different cancers,
12:55nobody had any idea.
12:57So I tried,
12:58I tried to figure this out and I
12:59created a head neck scorecard using
13:01some of the science and outcomes research.
13:03We had structural metrics.
13:05We had process metrics, pretreatment,
13:07treatment post treatment and then
13:08of course outcomes, oncologic,
13:10functional and patient reported
13:12outcome metrics.
13:13We created this big scorecard,
13:14reviewed all the registry data over.
13:16A few years divided things into
13:19surgical outcomes, medical outcomes.
13:21I'm not going into the details of this,
13:23and what I found was that it was
13:26actually hard to get the data.
13:30And 2nd,
13:31we didn't really have national
13:32benchmarks upon which to compare our
13:35data with the country as a whole.
13:37We were sort of in the realm of expert
13:40opinion to say how we were doing.
13:45So it was under that idea that I began
13:47the first stage of research in measuring
13:50quality and head and neck cancer.
13:53So these were some of the
13:55areas of interest that I had.
13:57I started first with volumes and outcomes,
14:00a structural metric. I didn't spend
14:02much too much time here because.
14:06It's been published and researched
14:07and sort of known to death.
14:09Higher volumes lead to better outcomes.
14:12It's been known for years and every aspect,
14:15probably of medicine definitely surges,
14:18definitely surgery.
14:19Here's a few studies published 2009, 2010.
14:23And I was surprised to see even
14:24in this last studies, 2022,
14:26people are still publishing the
14:28volume outcomes relationship studies.
14:30That was pretty well known,
14:31but I didn't want to be left in the lurch.
14:34So I looked at this as well.
14:35I said well, how about.
14:36I tried to look at Connecticut.
14:38Turns out the NCDB won't give you
14:39the data for Connecticut because
14:40there's not that many hospitals
14:42and you could probably identify
14:43which hospital is which one of you
14:45were trying to do the research.
14:47For example,
14:47I don't think there's too many
14:49hospitals in Connecticut that do 200.
14:51Analytic head and neck cancer cases,
14:54but I was able to get New England data
14:56and so we looked at data for upper
15:00aerodigestive tract cancers and we
15:02looked at average case volume by hospital,
15:04which was about 26 cases per year with a
15:09range of 1 case to 213 cases per year.
15:12And we did our standard multivariate
15:14analysis controlling for age,
15:16stage grade,
15:17comorbidity,
15:18insurance status and we found that
15:20worse survival was associated with.
15:23Treatments at a facility seeing
15:25less than 50 cases per year compared
15:27to greater than 50 cases,
15:28a pretty pretty impressive hazard ratio.
15:32Not new, not novel. But I wanted to know.
15:35I'm Canadian. I'm in New England.
15:36How are we doing?
15:38And that that gave me the the answer.
15:41So then I went on to a series of studies
15:43looking at timeliness in cancer care.
15:44So why is timeliness important?
15:46Well,
15:47timely diagnosis and treatment is
15:49associated with improved perceived quality
15:52of care and lower patient anxiety.
15:54So that's important.
15:55And then one could say that
15:58assuming the disease progresses
16:00while waiting for treatment,
16:02delays may result in more extensive
16:04treatment and possibly increased costs.
16:07And 3rd,
16:07the impact of treatment
16:09timeliness for cancer on true.
16:11Health outcomes like patient
16:13reported outcomes,
16:14functional outcomes,
16:15even survival a little bit less clear
16:17and this is going back you know 10 or 15,
16:19about 10 years when I was
16:21doing this research originally.
16:22So we started a number of studies.
16:24First we looked at treatment
16:26delays in oral cavity,
16:27squamous cell cancer across the country.
16:30We had three objectives in this
16:32study and this study was presented
16:33at the American Head Neck Society
16:35meeting as one of the best papers
16:37at the meeting and we looked at
16:39national trends in treatment time.
16:41Tables in patients with oral cavity cancer.
16:43So just get some benchmarks.
16:45How is the country doing?
16:47How can we know how we're doing at
16:48Yale or you're doing wherever you are?
16:50If we don't know what a benchmark is,
16:53then we wanted to identify factors
16:55associated with extended treatment
16:57time intervals and then to determine
16:59if delays in treatment intervals
17:01were associated with survival.
17:02These last two were more secondary outcomes.
17:04Really my main purpose of this was
17:06just to see how is the country doing.
17:09So we went through our standard.
17:11Inclusion exclusion criteria of
17:13oral cavity squamous cell carcinoma
17:15treated over a 14 year period using
17:17the National Cancer database and
17:19ended with about 6000 patients in the
17:21final cohort with oral cavity cancer.
17:23These are the time intervals that
17:26we measured diagnosis to surgery was
17:29diagnosis to treatment initiation.
17:31Surgery to RT start was what it says
17:33and then RT start to RT radiation
17:36start to radiation end was was at
17:38its was as it says and then we had
17:40total treatment package which is
17:42from when you started treatment to
17:44ending treatment and then diagnosis
17:46to treatment ends.
17:47So these are these are the intervals
17:49that we used and this is what we found.
17:53These are box plots you can hear
17:55you see huge variation across the
17:57country in every single one of
18:00these treatment intervals.
18:02As a quality researcher,
18:03if you are when you'll know,
18:04or any researcher really would know,
18:06that the wider these box plots,
18:09the greater variation and a huge
18:11opportunity for quality improvement
18:13is to shrink these box plots.
18:16We looked at survival.
18:18Again, secondary outcome,
18:20probably not the best database
18:22to look at survival.
18:23Plus we only had overall survival.
18:26What we found was that radiation
18:28duration if you were in the median
18:30or below versus the 4th quartile was
18:32significantly associated with worse survival,
18:35meaning treatment breaks during radiation.
18:37Not new,
18:38not novel but but that's what we found.
18:40And look at this survival curve.
18:42This is RT duration.
18:44You compare this to the ECOG 3311
18:47and the keynote O48 survival curves,
18:49which I flashed up there.
18:50I think this is a little bit more impressive.
18:53Not as great a study but so many
18:56more problems with it but I just do
18:59that for a visualization and these
19:01were this is what I thought was more
19:03more interesting though was the now
19:06finally we had time intervals across
19:08the country at NCNC DBCC accredited
19:11hospitals for for median times to
19:14treatment and and this is what we're at.
19:17We did this a lot and we did
19:19it for oral cavity.
19:20We did it for oral pharynx
19:22treated with surgery.
19:23We did it with oropharynx
19:27treated nonsurgically.
19:28We did it for hypopharynx cancer.
19:31Did it for salivary cancer and
19:34we established benchmarks.
19:35We we had box plots for all of these.
19:37We looked at survival outcomes
19:39for all of these.
19:40So I told you about oral cavity already.
19:42Here's oral pharynx,
19:43look at the the wide variety in
19:45the whole huge variation in treatment for
19:48oral fairness cancer treated surgically.
19:50Here's a survival curve.
19:52This was diagnosis to treatment end.
19:54We're seeing big differences if you were
19:57delayed versus not delayed and this was
19:59the same for oral pharynx cancer treated.
20:01Nonsurgically multivariate analysis
20:03controlling for all the standard factors
20:06that we always control for hypopharynx,
20:09cancer, salivary cancer, it goes on.
20:11So my point here is that analysis of
20:14variation in treatment time intervals really
20:17can identify opportunities for us to improve.
20:212nd, we found that there's a number
20:23of factors associated with delays.
20:24I didn't show you that that data here,
20:27but these can also are are often related
20:31to access and coordination of care.
20:33And so the third finding here was that
20:36every effort should be made to prevent
20:38prevent radiation treatment breaks,
20:40because in every single one of those
20:42there was a significant association
20:44with overall survival,
20:45meaning worse if you had treatment
20:48breaks or extended radiation.
20:49And this matters to patients, too.
20:51Here's a patient with an oral cavity cancer
20:54diagnosed elsewhere delayed getting in,
20:57starting with an oral cavity cancer,
20:59but eventually,
20:59when he came to treatment,
21:01extending through his skin,
21:04through the mandible.
21:06Original CT scan,
21:07no mandible invasion and here we are
21:10taking this cancer out mandibulectomy
21:13composite SO4 mouth mandible skin,
21:16using the fibula to reconstruct
21:18with a plate and using the skin
21:21on the outside to reconstruct so.
21:24Cancer is out, but at what cost?
21:26Because of delay in diagnosis,
21:29delay in treatment,
21:29delay in diagnosis is a whole
21:31other discussion,
21:32which I don't talk about and
21:35I don't really study.
21:39The Commission on cancer has
21:41just in March of this year,
21:44very recently finally put in their
21:45first head and neck oncology quality
21:48metric that's going to be measured at
21:50every single ACS accredited hospital
21:52across the country and that's time to
21:55initiation of post operative radiation.
21:57This my study didn't really
21:59show much of an association.
22:01There's tons of studies focused using
22:03that as their primary outcome that
22:05have shown that 42 weeks, 42 days.
22:07Is a big cut off to affect survival to start
22:10radiation after head and neck cancer surgery.
22:13So this is a new quality metric
22:15that we're all being measured on,
22:16just so you know. OK.
22:19So then did volume outcomes,
22:22did timeliness.
22:23So then I wanted to look at surgical quality.
22:25How do we measure that?
22:26How do we benchmark this?
22:27How do I know if I'm doing a good job?
22:29And so we started looking
22:31at a number of things.
22:32So this was Ben Judson
22:33was a lead author on this,
22:34but our whole team was involved
22:36where we look, tried to identify.
22:39Thresholds for lymph node yields.
22:42For oral cavity cancer,
22:43so our objective in this study was to
22:46determine lymph node yield threshold
22:48and oral cavity squamous cell
22:49cancer that might impact survival.
22:51This was a very interesting study because
22:55we used the NCDB to establish those
22:58yields and then validated it in SEER.
23:01And.
23:02To find the threshold,
23:04we looked at the clinically end zero oral
23:07cavity cancers versus the clinically
23:10and positive oral cavity cancers.
23:12And.
23:15Looked at different.
23:18Hazard ratios based on number of
23:20the lymph node yield basically and
23:23after regressive statistical models,
23:25we basically eventually found
23:27that there was a difference at
23:2916 lymph nodes for N0 neck N,
23:310 lymph node yields and 26
23:34for and positive lymph nodes.
23:36And after adjusting for all sorts of factors,
23:39we found a survival benefit,
23:41overall survival and cause specific
23:42survival based on lymph node yields.
23:45This is using SEER data so
23:46we finally had disease.
23:48Specific survival,
23:48which we don't have in NCDB.
23:50So again,
23:51if you look at these survival curves,
23:53I think they're quite.
23:56Not provoking.
23:58This study came out at the exact
24:00same maybe like a three months
24:02after our study and this looked
24:04at all of head neck cancer and it
24:06looked at used 18 lymph nodes,
24:08so kind of an arbitrary number they picked.
24:10To be honest based on some other
24:12single institution studies they
24:14didn't use the same statistical
24:15rigor that we use in oral cavity
24:17but it applied to all head neck
24:18cancer and so it was more applicable
24:20in 18 has become the big number
24:22for head and neck cancer partly
24:24based on this study but again.
24:26Think what you will just have a #18
24:29lymph nodes. Big survival difference.
24:32So that's lymph node yield.
24:33Then we wanted to look at positive
24:35margin rates and this is another
24:37study that we did trying to
24:38again figure out positive margin
24:40rates in this country for oral
24:41cavity, squamous cell cancer,
24:43what is the baseline?
24:45And again we found incidence of
24:47positive margins based on this study
24:49and we did look at the volume outcomes
24:52relationship and the facility setting.
24:54And if you were to academic center,
24:55your risk of having positive
24:57margins was significantly less.
24:5920 cases seem to be a good threshold,
25:01but again you see a lot of variety.
25:03Variation across the country.
25:05We also looked at transoral robotic surgery.
25:09So the transoral robotic surgery is FDA
25:12approved for T1T2 oropharynx cancers.
25:14And what I wanted to know was well
25:18what is the positive margin rate
25:21nationally and then what factors are
25:23associated with positive margin.
25:24So we did our standard inclusion
25:26exclusion criteria.
25:27We found about 2600 patients who
25:29had a transoral robotic surgery for
25:31oropharynx squamous cell carcinoma
25:33and what did we find overall.
25:36A 17% positive margin rate in this
25:38country and it varied by T stage
25:43T1T2T3T4. You can see that all here.
25:47We also found factors
25:48associated with positive margin,
25:49specifically Lymphovascular invasion,
25:51T classification and again facility
25:54volume of patients treated at high
25:56volume centers were less likely to have.
25:59Positive margin,
26:00you can see academic centers versus
26:03non academic and high volume versus
26:06high volume versus low volume.
26:08So in the, this is interesting because
26:11in the year since FDA approval,
26:13what we found is that the positive
26:15margin rate for towards the 17%.
26:16But if you look at the ECOG 3311 trial,
26:20the positive margin rate of
26:22credentials academic surgeons was 3.3%.
26:25And if you look at pooled data
26:28from 3 clinical trials looking at
26:31academic single institution studies
26:32or even a systematic reviews,
26:34we're looking at significantly
26:36less than the real world.
26:38Positive margin rate,
26:39so there is and then this study
26:42also showed a linear association
26:45between positive margins with T32
26:47and T4 tumors greater than 28%.
26:50So these patients are the ones that
26:52are going on to chemotherapy and
26:54radiation and probably had no benefit
26:57from transoral robotic surgery based
26:59on their current treatment paradigms.
27:01We did this for parotid cancer,
27:03I won't get into those details.
27:05And then this is a interesting
27:07study that combined margin rates
27:09nationally and lymph node yields and
27:12what they looked at was treatment at
27:15the try identify that the treatment
27:17at hospitals that attain a high
27:19rate of negative margins.
27:20So if you're if you have if you're
27:22a hospital that's that's high rate
27:24of negative margins and lymph
27:25node yields of more than 18.
27:27They found that there is a
27:29significant association with
27:30improved survival if you did.
27:31Both of those and these predicted
27:33outcomes independent of those generally
27:35modifiable characteristics including
27:37the volume outcomes relationship.
27:40These were independents of volume of
27:43the hospital but only 105 hospitals
27:46out of 1000 in in the country achieve
27:50negative margins in 90% or more patients and.
27:57If you look at lymph node yields
27:58of greater than 18,
27:59only 199 hospitals out of 1000 consistently
28:05achieved lymph node yields of 18 or more,
28:08meaning in an 80% or more of that's the case.
28:10We're not looking for 100%,
28:11you know,
28:1280% of cases having 18 or more lymph nodes.
28:16So in contrast to the traditional emphasis
28:18on the volume outcomes relationship,
28:20this showed that negative margins
28:22and lymph node yields can actually
28:24neutralize the effect of hospital volume.
28:29OK. So then another area that I want
28:31to look into was guideline adherence
28:33and this it came up because of a study
28:36published in 2009 from MD Anderson.
28:39They looked at 107 patients who came
28:42to them for second opinions with
28:44persistent or recurrent disease.
28:46So they've been treated elsewhere.
28:47They had persistent or recurrent disease and
28:50they wanted to look at well what's going on.
28:52What they found was that 43% of
28:54patients had NCCN guideline non
28:56compliance is the term that they used.
28:59And they try to find some
29:00factor associated with it,
29:02even the specific referring physician,
29:04the type of physician,
29:05the subset of disease,
29:06insurance status, age, sex.
29:07And there was nothing that was
29:10significantly associated with why someone
29:12would get NCCN noncompliant treatment.
29:15So I'm in the same vein.
29:18As you can see my theme here is,
29:20well, how's the country doing?
29:21You know we know how single institutions
29:23are doing, but how about the country?
29:24What's the benchmark?
29:25What are we aiming for?
29:27So we undertook a study to look
29:30at national NCCN guideline,
29:32not non adherence rates.
29:33That was the main objective.
29:35We also wanted to look at associations
29:38between non adherence and survival.
29:40In head neck cancer,
29:41we wanted to look at reasons,
29:43reasons for non non adherence
29:45and then factors associated with
29:47non adherence as secondary aims.
29:49So this is our we started with 375,000
29:54patients treated from 2004 to 2013.
29:58And we tried to figure out which ones
30:00were not adhering and which ones
30:02were adherent to NCCN guidelines
30:04and not going into the details.
30:06This is how we define non adherence by site.
30:09It's just look at the NCCN guidelines.
30:11It's actually pretty easy and they
30:14were surprisingly consistent in
30:15all these areas throughout the
30:17entire years of the study.
30:19We've got the guidelines going back to when
30:22this study started actually from the NCCN.
30:25And these are the numbers.
30:26So if you go to 2004 non adherence
30:28rates throughout this country,
30:30it was 30% for head and neck
30:32squamous cell carcinoma by 2013,
30:34we're still looking at almost
30:361/4 of patients are getting non
30:39adherent guideline non adherent
30:41care within this country.
30:43We looked at his by sub site
30:46for oral cavity cancer,
30:4746% of patients are getting non
30:50adherent care and oral pharynx,
30:52much lower hypopharynx,
30:54larynx and sinuses here.
30:57And then we looked at survival.
30:59So if you have non adherent care
31:01versus adherent care to NCCN
31:03guidelines we found a significant
31:05association with overall survival.
31:07Again a multivariate analysis
31:10controlling for all the typical factors.
31:13And then we looked at factors associated
31:15with guideline not adherence,
31:17black race, age over 65,
31:19comorbidity, non private insurance,
31:21higher T stage and then being
31:24treated and non academic facility.
31:27So I like this quote because it
31:30really talks about highlights why
31:32this is so important when clinicians,
31:34clinicians already know the information
31:36contained in guidelines and this
31:38was years ago, this was 1999,
31:40now they're so accessible.
31:41So I'm going to extend this to
31:42say when physical clinicians.
31:43Know how to get the guidelines
31:45within minutes at their fingertips.
31:47Those clinicians who want to improve
31:50quality need to redirect their
31:52efforts to identify the barriers.
31:54It's not knowledge that stand
31:55in the way of behavior change.
31:57So by addressing the barriers that
32:00prevent adherence to NCCN guidelines
32:03and their treatment protocols,
32:05I think there is a great opportunity
32:08for us to improve survival.
32:10I would even say a responsibility.
32:12Final area of research I wanted to look
32:15at and I haven't done much in this yet.
32:17So I'm going to quote a a different
32:20study by Bevin you who's the chair
32:23at Minnesota University of Minnesota
32:25and he was actually a Robert Wood
32:28Johnson scholar here at Yale and very
32:30well known and outcomes research.
32:32And that cancer he just wanted,
32:34this was a spoke to him about his
32:35very controversial study when he
32:37published this because it was within
32:38his own network of patients that
32:40he was seeing looking at patients
32:41who were treated with radiation.
32:43Therapy at the academic center versus
32:45the non academic center and found
32:47significant differences in five
32:48year survival on there were similar
32:50rates of treatment completion,
32:52similar rates of treatment breaks,
32:54more advanced cancers at the academic center.
32:57Multivariate analysis really you
32:59know statistically rigorous study
33:02couldn't say why but there was.
33:05OK.
33:06So the last thing I want to talk
33:08about is really this concept of bench
33:11to bedside for the patient centered
33:13outcomes researcher and that's what
33:15I've been doing over the last year or so.
33:19What I wanted to do now is take all
33:21this national benchmark data we had.
33:23We knew what good quality was or
33:24at least what the standard of care
33:26was across the country.
33:27I said, hey, how,
33:28how is Yale New Haven Health system doing?
33:31And so this is what we did, we looked.
33:36At tumor registry data from Yale,
33:39New Haven Hospital, Bridgeport Hospital,
33:41Greenwich Hospital, L&amp;M and Westerly.
33:43So our entire network and we got
33:45all the analytic oral cavity cases
33:47from 2012 to 2018.
33:48And we looked at a number of
33:50these quality metrics for which we
33:52now had national benchmarks.
33:54And we looked at the positive margin rate.
33:55We looked at lymph node yields
33:58greater than or equal to 16 adherence
34:00rates to NCCN guidelines and time to
34:03adjuvant therapy within six weeks.
34:05So we started with 500 patients
34:07and these were the three groups of
34:09patients that were treated academical
34:11surgery and radiation or all their
34:13treatment if there was no radiation
34:15at the academic only center,
34:16Community Center being the opposite
34:18end and then the combined group.
34:20So they had in this combined group,
34:22it was always surgery at the main
34:24center and radiation elsewhere.
34:25For some reason it didn't go the other way.
34:28And these were our key findings.
34:31There was a higher positive surgical
34:34margin rate within our health
34:36system at the Community hospitals,
34:3812% versus 2.5%.
34:39There was a lower likelihood to
34:41meet that bottom lymph node yield
34:43threshold of 16 lymph nodes and a
34:46neck dissection at the community,
34:4859% versus 90%.
34:49There were decreased rates of
34:51adherence to NCC and guidelines in
34:53other hospitals within our Network,
34:5576 versus 86% and adjuvant radiation
34:58therapy within six weeks it was the same.
35:01If you were all academic or all community,
35:03but if you fragmented your care,
35:05it was significantly less.
35:0722% of patients, when fragmenting their care,
35:10received radiation therapy within six weeks.
35:14So it sounded negative originally,
35:15but then we said, hey, well,
35:16we have national benchmarks.
35:18How are we doing?
35:18It turns out that we're actually
35:19doing pretty good within the system.
35:21The, our numbers are on par with the nation.
35:23We're not any worse than the national
35:26benchmarks in our community hospitals.
35:28If you look at the national positive
35:29margin rates based on our studies,
35:31oral cavity cancer is almost 13%.
35:33Our community partners are 12%.
35:37I mean our academic center is 2.5%.
35:39So that's a significant difference,
35:40but but the system is.
35:44Not doing horrible.
35:46There's a lot of reasons that there could,
35:48you know, there could be differences.
35:50For example, availability of frozen section,
35:52access to advanced reconstructive surgery,
35:54we can get more aggressive on the resection,
35:58constant communication between
35:59pathologists and surgeons,
36:01presence of trainees,
36:02perhaps higher case volumes,
36:04how margins are taken by surgeons or
36:06how they're assessed by pathologists.
36:08Countless reasons this for this,
36:11all theoretical.
36:13The other point about positive margins
36:15other than affecting survival is
36:18that it can often lead to unnecessary
36:20escalation of cares like that oropharynx
36:22cancer case that I showed you.
36:24But in this cohort of patients
36:26within our system,
36:27two patients in the Community only
36:30group received adjuvant chemotherapy
36:32plus radiation for early stage disease.
36:34So stage 1/2 cancer positive margin
36:36went on to chemotherapy and radiation.
36:38There's no LINOP and I in those cases.
36:41And when you look at non adherence rates.
36:44Again nationally for all of cancer,
36:4726% of patients for head neck
36:48cancer nationally are non adherence
36:50with guidelines for oral cancer.
36:52If you'll recall based on our
36:54previous studies,
36:5546% of patients are non adherent
36:57in our community.
36:59Yale New Haven Health System is doing
37:00pretty good when you look at national data,
37:02much better than than the
37:04country for oral cavity cancer.
37:05But if you look at the academic center,
37:07we see a significant difference
37:09and again no one looking for
37:11100% adherence to guidelines,
37:13that's not possible, a lot of reasons,
37:15sometimes you can't.
37:15Adherence to guidelines,
37:17but we're looking to benchmark against
37:19the country and against each other.
37:22So.
37:22We looked at lymph node yields
37:24and this is where things got a
37:26little bit more interesting.
37:27Nationally,
37:27greater than or equal to 16 lymph nodes,
37:3070, we used 18.
37:32That's what we have for national data.
37:3470% of patients have more than 18
37:36lymph nodes in their neck dissections.
37:39In our community,
37:4059% and in the academic center,
37:4290%. And you can see the huge
37:45variation as well everywhere really.
37:48Again, there's multiple reasons for this.
37:49How pathologists count limp,
37:51it's not all surgical related,
37:52how pathologists count lymph nodes,
37:54but maybe also the quality of the
37:56neck dissection and this these
37:58can sell a lot of these can be
38:00modifiable treatment factors.
38:01So in this study,
38:02we found that treatment of oral
38:04cavity cancer at Community facilities
38:06within our system may increase the
38:08likelihood of positive margins,
38:09lower likelihood of adequate lymph
38:11node yields and decrease adherence.
38:12NCCN guidelines.
38:13Not only that, but if you fragment your care,
38:16you can have a lower likelihood
38:19of achieving radiation initiation
38:21within six weeks.
38:22A lot of problems with this study.
38:24I think it's still very informative
38:26and very useful to make us better,
38:28but you know,
38:30data were collected retrospectively.
38:32We found a lot of differences in process
38:34related metrics and other studies
38:36have shown association with survival.
38:39This study in and of itself
38:40did not show that there was
38:41any difference in survival,
38:42definitely not powered for this.
38:44It was not the main outcome and
38:46of course there are socio.
38:47I recognize there are socioeconomic and
38:49other factors that can impact patient
38:52care delivery and confound the results.
38:54For example,
38:55patients who are willing to
38:56travel to the academic centers may
38:58differ systematically from those
38:59who seek care closer to home.
39:04Since then, I'm not going to read all
39:06this and tell you this is historic
39:08data going back to 2012 to 2018.
39:11Smilo has, I think, been aware of this.
39:13The leadership has been aware of
39:15this and there's been a lot of
39:17things that have changed to improve
39:18quality even in the last five years.
39:20And I'd love to do this study and I
39:21will do the study a few years from now
39:23as well and see how much better we are.
39:25For example, oncology care pathways
39:27have been initiated.
39:28Community hospitals and disease teams
39:31have been better integrated at SMILOW.
39:34We are creating local disease teams,
39:36experts at community hospitals primarily
39:39dedicated to head and neck cancer care.
39:42Physicians at the academic Center have
39:45established clinics in surrounding
39:46care centers.
39:47We're trying to bring culture from
39:50one place to another as well and
39:53better integrating and aligning
39:54culture across our care centers.
39:57Of course,
39:57advancement in ER has helped a lot.
39:59This is going back to 2012
40:01when the ER was first.
40:03Initiated in the health system,
40:06we have more clinical trials and
40:08they're offered at care centers
40:09and there's a lot of other
40:11improvements that are ongoing.
40:13And I think delivering high quality
40:15care across the system for head and
40:17neck cancer is an important part of
40:20our systems growth and integration.
40:24So really what I think is that integrated
40:26health systems can leverage the strength
40:28of the academic center to figure out
40:31a way to disseminate best practices,
40:33to break those structural barriers
40:35within our network, to improve patient
40:37care and bring high quality care,
40:38whether it's closer to home
40:40or bring the patients here.
40:41But we can do this and dare I say
40:43we have a responsibility to do this,
40:45but how are we going to overcome
40:48these structural barriers?
40:49So bringing it back to the patient.
40:52You know, processing,
40:53quality care improves survival.
40:54I told you that.
40:56But quality care.
40:57It means more than just the
41:00process of care and survival.
41:02It means more than just being cancer free.
41:07Say
41:091234512345. Left and right.
41:13Just to the outside.
41:17Hi, this is **** Morris.
41:19You may have seen me on some of
41:21my Facebook or Internet posts.
41:23I had tongue cancer and doctor Mayer.
41:27Doctor Mehra operated and after
41:29the portion of the tongue that was
41:33cancerous was removed, he replaced
41:35it with a skin graft from my forearm.
41:39I'm now speaking to you 5 weeks to
41:41the day after surgery and I no longer
41:44feel like I have marbles in my mouth.
41:47I can speak clearly and
41:50it's entirely due to him.
41:52Yeah, Mike, partial Larry.
41:54Mack Larnax, see perfect was four months ago.
42:00And.
42:02I'm amazed that I have this voice.
42:06Go
42:101234567. Do you have a feeding tube?
42:13No. Do you have a tracheotomy? No.
42:18Ohh, I had my surgery three months ago.
42:25Free **** pie,
42:27I had my surgery
42:29on May 2nd by Doctor Merritt.
42:31So you can see these are patients
42:33who where function is important.
42:35It's not just that they're alive
42:37and they're cancer free and they
42:39had guideline compliant care,
42:40but you have a laryngectomy patient
42:42speaking hands free thanks to our
42:44speech and language pathologist.
42:45You've got patients who are in the
42:47public speaking after having significant
42:49portions of their tongue removed,
42:52but high quality reconstructions.
42:54Partial laryngectomy is for
42:56salvage which are not done at,
42:59certainly not at in low volume places,
43:01saving saving the larynx.
43:05So I hope what I've done now is
43:07you already knew that randomized
43:09control trials are great.
43:10They lead to practice changing innovations
43:12in the care of cancer patients.
43:14But I hope you can have a better.
43:17Respect maybe for quality
43:19of care research as well.
43:21We can have a remarkable impact
43:23on patient outcomes and I think
43:25it's necessary for the quote real
43:27science to mean anything to patients.
43:29You have basic science researchers who
43:31feed the clinical researchers within
43:33clinical researchers feed the patient
43:35centered outcomes researchers and that's
43:38what leads to better patient care.
43:40So Healthcare is remember this is
43:42getting the right care to the right
43:43patient at the right time every time.
43:45That's quality cancer care.
43:47Thank you.
43:56Thank you so much for both for the
43:58wonderful talk and the wonderful work.
43:59Are there any questions either
44:01here in the room or I know people
44:05are also potentially posting?
44:10Through the webinar.
44:14Yeah, Tommy.
44:18Open.
44:21COVID-19.
44:27I don't know yet is a short answer.
44:30I think that's it's an area of active study.
44:32There are some small scale
44:34publications on that,
44:35but I think we're going to know.
44:37Pretty soon, I can tell you our experience,
44:39but that's not based on the data.
44:40Maybe Doctor Burtness has
44:41more information about that,
44:43but I think we're going to
44:44know better soon enough.
44:48I mean I I think we saw a number of,
44:51I don't know that we have stats on it,
44:53but we saw a number of patients with
44:57considerable delays in diagnosis.
44:59So I have a I have a question.
45:02When we participate in,
45:03so dragging you back to the
45:05clinical trial side for a moment,
45:07but a really important consideration for
45:11surgically based trials is how do we set
45:15what the benchmark is for the right surgery.
45:19And I know you're participating in
45:21the Sentinel lymph node trial now,
45:24sort of how does that credentialing process?
45:28Sit with our own quality enhancement efforts.
45:34It's a good question.
45:35I think you're right.
45:36It's just exciting now to have
45:38surgeons involved in randomized
45:39control trials like ECOG 3311,
45:41the Sentinel node trial.
45:42It's actually great because it hasn't
45:44happened a lot in head neck cancer.
45:47You know, I think that for clinical
45:49trials it's different than patient these,
45:50these comparative effectiveness
45:52researchers because you need to know if
45:55the intervention is effective or not.
45:57You need high quality surgeons every
45:59time they need to be credentialed
46:01and the surgical quality metrics
46:03are being monitored constantly and
46:04I think you need that for clinical
46:06research and I'm glad it's there.
46:08But how you apply that to the
46:10real world I think is,
46:12you know at the second
46:13area of research interest.
46:14But I'm glad they're credentialed
46:16high quality surgeons during
46:18the clinical trials.
46:20We
46:20have a a question from Doctor
46:22Robinson in radiation oncology.
46:24How do you disseminate the
46:29these quality insights to?
46:32The lower volume centers and are there
46:35specific guidelines for which patients
46:37should be referred for high volume?
46:40Is it practical to refer everyone
46:43with less common cancers like head,
46:45neck cancer? And if not,
46:48how do you feed these?
46:52These care care approaches
46:54to the lower volume centers.
46:57That's a great question.
46:58I think it's $1,000,000 question.
47:00If we could do that honestly I
47:02think we could improve survival more
47:04than the next billion dollar drug
47:06basically and I think that's that's our
47:08challenge that's what we need to do.
47:10I don't know I think every system
47:12has it has a different approach.
47:14I think Yale New Haven Health System
47:15is going to be different than
47:17memorial is going to be different
47:18than pen is going to be different
47:20than any health system but.
47:21I don't know what the right answer is,
47:23but we need to figure it out.
47:28I guess I have a follow up question
47:31to that and and that is some of
47:34the hypothesis you had about what
47:37controlled let's say the rate of
47:40margin positivity had to do with.
47:45Frozen margin evaluation communication.
47:52Are are there components of that that
47:55could be enabled by telemedicine?
47:57So you know you referred to digital
47:59pathology once or twice in your talk,
48:01but are there ways that expert disease
48:06focused let's say pathology evaluation?
48:10Could be extended through telemedicine.
48:14Yeah, I think that would be,
48:16I think that's one way to bring expert
48:20level care across the system with
48:22telemedicine and I think pathology
48:25would be a great way to start actually.