FUTURE RESIDENT, DO NOT BECOME A RADIATION ONCOLOGIST!!!

[deleted18755]

This is really a simple issue and it's shocking to see all the sanctimonious talk on this thread.

Simply put: If being in major coastal city is very important to you, you probably shouldn't apply to this field unless you are an astounding candidate -- even people from top 10 programs have a lot of trouble landing there.

If you are flexible and okay living in the Midwest or South, then you'll be more than happy.

End of story.

This is probably the best summary but I think it should be modified to a bit to:

"If you are flexible and okay living (ANYPLACE) in the Midwest or South, then you'll be more than happy (AT LEAST FOR THE NOW).

I know somebody who has wanted to move to Charlotte, NC, which is actually a large and booming city recently, but there hasn't been a job there in years. He has actually applied to that job mentioned above in Boone, NC, which is a satellite 1.5-2 hours outside of Charlotte (and the closest airport) in a college town (Applechia State which appears to actually have major division I sports, a Home Deport, Barnes and Nobles, etc so not that rural). Apparently nothing else in the entire state of NC, which I was surprised to just look up and discover somehow has a population over 10 million people, other than in the extreme East, which is far away, or apparently at UNC but you have to be the chief at Harvard to get a job like that!

PS: RollTideRadOnc - I assume you're from Alabama? I actually also know somebody who has been trying to move to Alabama (closer to Atlanta not New Orleans but he might take anything) or Georgia for a few years but can't find a job. Apparently Birmingham is a tough market since the residents from that program like to stay in the area and he gave up on anywhere within an hour of Atlanta because it was cut-throat. Great guy so if you know of anything in the area please PM me and I'll forward you the contact (I'm almost certain he is still interested and might also be interested in whatever state is west or north of Alabama).

---

Members don't see this ad.

 

[Mandelin Rain]

People go into this specialty because it's low stress and high paying, particularly on a stress -to-dollar scale. That still exists. Where the expectation that you also get to live where ever you want came from is beyond me. It's not FP where if you can't find a practice to join in your desired area, you can hire an MA, rent a 3 room space, and hang a shingle. You want to do that in rad onc, you better have 6 million to risk in initial capital to start up.

And yes. Some one has to treat rural population. And yes, if we define rural as "undesirable" to the entire pool of rad onc doctors, the market will ultimately select the less desirable candidates for those jobs. That doesn't mean "bad doctor". It means you are competing with other incredibly intelligent and hard working people who may have been more qualified, or more connected, or more personable, or more positive, or less miserable that also desire the more desirable jobs. My suggestion if you find yourself in a locale you find undesirable, don't lament your fate on the internet while being a miserable human being. Take excellent care of your patients. Educate your referring docs. Educate your hospital administrators regarding the importance of new tech. Build your program. Be positive and make connections in your desired area so that when a job open there in a year or two and they call your references, they'll be so glowing they will have no choice but to hire you. This is not as easy as glibly accepting your fate with the learned helplessness of a turtle turned on it's shell, but it is doable.

 

[radiaterMike]

I agree that this is a great synopsis from Roll Tide. I would add that if you do not want to be in the middle of rural nowhere, you do not have to be. As you say- you may not land a job in a major coastal city (although certainly some new grads do- even from non 'Top 3' schools) but you also will not have to abandon civilization altogether.

OldKing - most from Ann Arbor are familiar with Appalachian State (as are most from East Lansing and Columbus)

 

[DrfluffyMD]

Actually being an desirable/undesirable candidate matters less and less as the job market becomes tighter while the “luck of the draw” becomes higher.

You can be the best darned chief Harvard ever has but if NYC needs 20 radonc and all those jobs are filled, you aren’t getting a NYC job that year period.

 

[psy3]

People go into this specialty because it's low stress and high paying, particularly on a stress -to-dollar scale. That still exists. Where the expectation that you also get to live where ever you want came from is beyond me. It's not FP where if you can't find a practice to join in your desired area, you can hire an MA, rent a 3 room space, and hang a shingle. You want to do that in rad onc, you better have 6 million to risk in initial capital to start up.

And yes. Some one has to treat rural population. And yes, if we define rural as "undesirable" to the entire pool of rad onc doctors, the market will ultimately select the less desirable candidates for those jobs. That doesn't mean "bad doctor". It means you are competing with other incredibly intelligent and hard working people who may have been more qualified, or more connected, or more personable, or more positive, or less miserable that also desire the more desirable jobs. My suggestion if you find yourself in a locale you find undesirable, don't lament your fate on the internet while being a miserable human being. Take excellent care of your patients. Educate your referring docs. Educate your hospital administrators regarding the importance of new tech. Build your program. Be positive and make connections in your desired area so that when a job open there in a year or two and they call your references, they'll be so glowing they will have no choice but to hire you. This is not as easy as glibly accepting your fate with the learned helplessness of a turtle turned on it's shell, but it is doable.

While someone has to treat rural patients, it doesn't mean that the care needs to be next door. Yes, given this country is so large, there are true remote places that really have hard time attracting physicians, especially specialists. That's why at least I wouldn't mind radonc programs opening in Arkansas, West Virginia, Coastal NC. If those residency can attracts people who want to or will accept to be around as attending, why not? All of us have seen this when we were applying for medical school, schools in plain states, etc focusing on finding personalities who can stay around through personal statements and interviews?? However, the current job crisis didn't happen because of development of those rural residency programs. Maybe there was some contribution. But, most of access supply came from expansions of already large programs, programs in large city, smaller hospitals in major city opening new residency, on and on. And the leadership allowed this citing maldistribution. And that is what's alarming for the people complaining about the current job market.

Also, regarding rural healthcare access. US is the only country that would put standalone Linac and infusion center (without inpatient medicine and surgical help to handle any active issues) in rural town with population of 10-20000. If you google map Canada, there is one cancer program (not even sure whether it includes radonc) between Toronto and Ottawa over 290 miles. And many if not all other advanced nations are wired that way, I would assume. So, this country somehow has all these centers in rural locations, pay for all the infrastructures, machines, and personnel which are not free, and the country as a whole has the absolute highest cost of healthcare that many of those underserved rural population cannot possibly afford.

So, yeah, in principal, nobody would argue that underserved population in rural area needs healthcare, I would questions how it is delivered in this country. But what I find more disturbing about your viewpoint is that you take the high road or even socialist view of "how these young ones have the sense of entitlement and deny equal and fair distribution of healthcare access," then you take the neoliberalistic claim that "oh, all those losers of competition of scoring high on Steps, publishing papers, getting into premier programs, etc, etc, just go to rural areas and work hard since you lost."

I think the lessons for medical students from this thread is that if the views of majority of current practicing mid or late career attending are like yours, ie "you lost to the competition of getting into top tier program and failed to make up through networking and cold calling, then you better put your head down and go serve rural area," in addition to the non-supportive academic leadership, then middle of the road med students really should think twice about getting into this field.

 

[scarbrtj]

Radiation Oncology at Thomas Jefferson University: A Specialty Emerges as a Department Evolves

"The first document, A Prospect for Radiation Therapy in the United States, was published in 1968. This widely respected and utilized document set the organizational structure for radiation oncology staffing of faculty and resources for upcoming practices across the United States. Dr Brady noted that because of a certificate of need was common in many states; government and insurance agencies used the blue book to determine the necessity for radiation therapy facilities. He recalled that many of the initial proposals were based on the United Kingdom system of care for which the committee recommended three types of cancer centres, a university centre treating both definitive and palliative cases, a second level centre treating fewer definitive cases and a third level centre treating palliative cases only."

For better or worse, this "rad onc caste system" mentality still exists today. Those poor, simpleton rural rad oncs

 

[TomatoMan]

I won't pretend to have the answer, but all these issues really need to be discussed more fully and openly by our leaders. I just want the discussion to happen.

100% agree. Honestly, we should try to assemble a coalition of some sort to collectively express the significant concerns practicing radiation oncologists have about the future of our specialty. Perhaps a completely new organization (which actually represents the interests of practicing rad-oncs, rather than hospital admins as ASTRO does) that ideally would have a seat at the table when new residency expansions are considered? I wonder if ACRO would be interested in something like this?

 

[OTN]

This is precisely the kind of issue ACRO should be addressing.

 

[medgator]

This is precisely the kind of issue ACRO should be addressing.

If anybody is going to the ACRO annual meeting next week, it would be great to hear back about what (if anything) is discussed regarding this issue.

 

[TomatoMan]

Randomly I clicked on one of the "similar threads" that show up below the active thread on the SDN webpage. It's from 2005...

"...rad onc is perhaps the most difficult specialty as there're only about 100 spots in the nation. it pays really well. best life style; work aroudn 30-45 hrs / wk! ur tx suprisingly to many, perhaps, will work well. u r not around radiation, contrary to popular belief, as it's the technicians that get near it. u r the mastermind of the whole rad onc tx team; u think, u make/adjust plans, and order others to follow them..."

Not sure if it was ever that cush, but my how things have changed...

how does the residency for a radiation oncologist work?

 

[seper]

It's like deciding to get married... 13 years later you suddenly find you spouse to be different in many ways.

 

[DrfluffyMD]

Randomly I clicked on one of the "similar threads" that show up below the active thread on the SDN webpage. It's from 2005...

"...rad onc is perhaps the most difficult specialty as there're only about 100 spots in the nation. it pays really well. best life style; work aroudn 30-45 hrs / wk! ur tx suprisingly to many, perhaps, will work well. u r not around radiation, contrary to popular belief, as it's the technicians that get near it. u r the mastermind of the whole rad onc tx team; u think, u make/adjust plans, and order others to follow them..."

Not sure if it was ever that cush, but my how things have changed...

how does the residency for a radiation oncologist work?

Pathology will be just as cush and competitive as radonc if there are 1/10th of its current spot. Overtraining will destroy any field.

 

[DrfluffyMD]

It's like deciding to get married... 13 years later you suddenly find you spouse to be different in many ways.

People who did a MDPHD and then radonc then finally found themselves unable to find good jobs in a drastically changed field is more akim to start a long distance marriage and then finally living together 13 years later and realizing it isn’t what it’s cracked up to be.

 

[MegaVoltagePhoton]

People who did a MDPHD and then radonc then finally found themselves unable to find good jobs in a drastically changed field is more akim to start a long distance marriage and then finally living together 13 years later and realizing it isn’t what it’s cracked up to be.

Are you basing this on your experience as a RadOnc resident?

 

[DrfluffyMD]

Are you basing this on your experience as a RadOnc resident?

Absolutely not. I am not a radonc. I am not sure where whether I am a radonc or not fit into this discussion because my statememt is a generalized one that applies to all careers where substantitial initial effort is invested and the return (or percieved return) is subpar.

For the OP, the return is subpar. As an outside observer, a sub 300k compensation in a highly nondesirable area for a new grad when the said grad had credentials competitive for DR, ortho, and competitive IM residencies seem to be quite subpar. It doesn’t take a radonc to see this issue if what the OP is presenting is mostly factual, which is corroborated by multiple posters.

 

[Palex80]

Pathology will be just as cush and competitive as radonc if there are 1/10th of its current spot. Overtraining will destroy any field.

Pathology actually may be doomed. Algorithms may replace pathologist in the next decade. Not all of them, but certainly many of them.

 

[scarbrtj]

Pathology actually may be doomed. Algorithms may replace pathologist in the next decade. Not all of them, but certainly many of them.

Google AI detects breast cancer better than pathologists - Pharmaphorum

 

[RickyScott]

The way DVHS/volumes/constraints are being standardized in some large systems- that can apply to some extent to radiation.

 

[psy3]

The way DVHS/volumes/constraints are being standardized in some large systems- that can apply to some extent to radiation.

It's already here. I think Google's venture (who acquired Deepmind) with UK NHS cited below hasn't panned out, but I'm sure more is coming.
Applying machine learning to radiotherapy planning for head & neck cancer | DeepMind

 

[OTN]

It's already here. I think Google's venture (who acquired Deepmind) with UK NHS cited below hasn't panned out, but I'm sure more is coming.
Applying machine learning to radiotherapy planning for head & neck cancer | DeepMind

Until a machine can correctly - with perfect accuracy- contour even the optic chiasm I'm not going to hold my breath. I haven't yet seen an automated contouring program that can even do that without substantial MD adjusting of volumes.

 

[maruchan]

Until a machine can correctly - with perfect accuracy- contour even the optic chiasm I'm not going to hold my breath. I haven't yet seen an automated contouring program that can even do that without substantial MD adjusting of volumes.

I'd caution against that logic. The progression of AI/machine learning isn't at all linear, recent advances have completely changed the game. Those advances haven't really been applied to rad onc yet though some are doing early work.

Automated contours are probably a few years away, but the obstacle is having high quality, large datasets of contours. The level of AI needed to draw good contours is already there.

I don't think it'll affect the job market too much though as contouring is a relatively small proportion of a rad onc's time and is rarely the limiting factor in the number of patients you see. Plus you still need a radiation oncologist to decide what to cover, make judgment calls about PTV coverage vs OAR doses, and in general do everything other than drawing shapes on a screen.

 

[medgator]

. As an outside observer, a sub 300k compensation in a highly nondesirable area for a new grad when the said grad had credentials competitive for DR, ortho, and competitive IM residencies seem to be quite subpar. .

Ummm where is that happening? You're going to see sub-300 in reasonably desirable areas only afaik.

Rural town USA is still going to be starting 350-400K minimum, if not more, depending on experience. I know a rural position that paid $600k+ a few years ago that a newer grad from my program took before he moved to be closer to family.

There are still well-paid positions out there depending on what type of location compromise you are willing to make and how busy you are willing to be

 

[scarbrtj]

Plus you still need a radiation oncologist to decide what to cover, make judgment calls about PTV coverage vs OAR doses, and in general do everything other than drawing shapes on a screen.

actually also perfectly suited to machine learning

 

[DrfluffyMD]

Again, not a radiation oncologist but pretty familiar with machine learning. I think the big threat to radonc isn’t a complete machine replacement, but rather a massive increase in efficiency due to a slick, integrated workflow with clinical decision support.

Radonc and oncology’s reliance on big data and high quality trials actually can make it easier for a high quality clinical decision support system to emerge faster.

Suddenly you maybe faced with a few big name jn academia doing all the cerebral work and radoncs in satelite campuses doing machine babysitting. God forbid they change the babysitting role to allow NP to do that.

 

[medgator]

actually also perfectly suited to machine learning

When we see competent autonomous vehicles for all situations,, then I'll start to get worried about what else AI and "machine learning" can do.

Right now, aggressive drivers at 4-way stop makes Google's Lexus RX go crazy

Stop lights, sunsets, junctions are tough work for Google's robo-cars

 

[Haybrant]

this really is one of the most important things in the whole thread from oldking above

"I know somebody who has wanted to move to Charlotte, NC, which is actually a large and booming city recently, but there hasn't been a job there in years. He has actually applied to that job mentioned above in Boone, NC, which is a satellite 1.5-2 hours outside of Charlotte (and the closest airport) in a college town (Applechia State which appears to actually have major division I sports, a Home Deport, Barnes and Nobles, etc so not that rural). Apparently nothing else in the entire state of NC, which I was surprised to just look up and discover somehow has a population over 10 million people, other than in the extreme East, which is far away, or apparently at UNC but you have to be the chief at Harvard to get a job like that!"

Let this sink in med students bc this is the reality. I know someone from top 5 program that needed to move to alabama bc his wife would be working there, he cant find anything within 2 hours distance so right now he isnt working. Let that sink in. You essentially have to be geography agnostic 100% for rad onc.

 

[kristofer]

I guess I'm shocked by total disregard of providing access to cancer care to non-metropolitan human beings because you happen to find where they choose to live disgusting (perhaps, you find they themselves disgusting?) or there is no immediate access to pelvic MRI.

If you can't figure out how to help cancer patients without an EUS or pelvic MRI, while working with perfect referring docs who do everything by the text book, perhaps we do need fellowships in this field.

I don't understand the arrogance and entitlement of newly minted radiation oncologists. That somehow working in a small town is beneath you, despite the market pointing to the opposite. We get it, you had a high Step 1 score 8 years ago, and published two retrospective reviews that no one read or cared about. Here's the world.

I'm glad to see that there is at least one voice of reason in this thread. I have been constantly annoyed by the ignorant attitude of many in this field that working in anything other than a major coastal city is beneath them. It really is sickening to me. As someone who struggled to get into this field despite reasonable qualifications who has a genuine interest in serving in an "undesirable area," I am extremely annoyed by the costal academic and socioeconomic snobbery that seems to inundate this field. There are myriad number of residents and attendings who went into this field purely for lifestyle reasons, and as result we have a "maldistribution" problem. The field has snubbed and discouraged otherwise good applicants who would gladly work in rural Kansas or Alabama. And instead we have an oversupply of people who are concerned with nonsense such as variety of restaurants and think that every rural white person is a neo Nazi who is out to get them. They want to self-segregate in large cities, send their kids to homogeneous private schools, and do everything possible to avoid rank and file middle americans or entertain the idea that culture there is anything other than deplorable. Who's the real racist? Not every patient can be treated in a major academic center on the coast, and there are "very fine" people deserving of care in "deplorable" places. And for what it's worth, I just visited a major costal city, typically cited as one of the most "desirable" cities to live in and witnessed a homeless man masturbating on the sidewalk, human feces in alleyways, severe overcrowding, unreasonable costs of living, and general societal devolution. You can have it.

 

[Palex80]

Decision making will be greatly added / substituted by machines in the future. And this goes for all sectors in medicine, not only radiation oncology.

Medical oncology is also going to get a hit, when machine algorithms are going to tell you what chemo-/immunotherapy regime you are to use, after a complete sequence of the cancer genome has been performed and other already documented patient factors like comorbidity, age, etc are taken into account. Big data will help to guide the right decision.
The same pattern will also be visible in other disciplines too, like infectiology or pharmacology. It's already happening right now, actually. I remember having to look up what medication I was prescribing to a patient and if any potential interactions between the drugs were to be expected. Nowadays, we prescribe electronically and the system will tell me if I am prescribing two drugs that may influence each other in any way. A system like that pretty much eliminates something like 30% of the workload of a clinical pharmacologist. Those guys used to come to ward rounds once per week in the past and point out at interactions of prescribed drugs, they don't do that anymore.


Surgical disciplines are going to get less of a hit, since autonomous robotic surgery is not to be expected soon and with robotic surgery the workload (as in hours spent in the OR) for a surgeon actually increased.
And some specialities still perform quite delicate manual exams, which cannot be completely overtaken by machines, like opthalmology.

Automatic contouring may happen at some point, but probably only for straightforward cases. It's going to hurt radiology first before it hurts radiation oncology. A typical scenario, where a machine can probably perform better than a radiologist would be lung nodule evaluation in chest CTs. A machine can be "fed" with a lot of CTs and will at some point be able to detect, classify and compare nodules a lot faster and more reliable than a radiologist. Automatic volumetric comparisons of repeat CTs of the same patient should also be more precise than a radiologist's judgement.

 

[MegaVoltagePhoton]

Absolutely not. I am not a radonc. I am not sure where whether I am a radonc or not fit into this discussion because my statememt is a generalized one that applies to all careers where substantitial initial effort is invested and the return (or percieved return) is subpar.

For the OP, the return is subpar. As an outside observer, a sub 300k compensation in a highly nondesirable area for a new grad when the said grad had credentials competitive for DR, ortho, and competitive IM residencies seem to be quite subpar. It doesn’t take a radonc to see this issue if what the OP is presenting is mostly factual, which is corroborated by multiple posters.

It is corroborated by some posters. The job market for rad onc is far from perfect, and I agree that people should enter the field with a realistic expectation of geographic restrictions. That said, what makes you think the kind of person who wants to go into rad onc wants to go into ortho? Or even diagnostic radiology? The fields are vastly different in innumerable ways. Similarly, what exactly are these magic hospitalist jobs in SF and NYC? The med students should also realize that being a hospitalist =/= being House, MD, it means memorizing the CIWA scale, treating scabies, managing postop ortho patients who have mildly elevated blood sugars, etc. At the end of the day, the rad onc in coastal NC is presumably still a treating cancer. If treating cancer is not that important to an applicant, then that applicant is probably not the best fit for rad onc anyway. These are also important tradeoffs to consider.

 

[jupiterianvibe]

Pathology will be just as cush and competitive as radonc if there are 1/10th of its current spot. Overtraining will destroy any field.

Overtraining isn't just bad because it destroys your field's power, income, and prestige. It's bad most of all because it lowers the standards to the point where patients can actually be harmed.

I have run into too many marginal or dangerous FMG pathologists to know that having an oversupply problem is bad for patient care, first and foremost.

You guys are pretty smart as a whole now, but consider that if radonc keeps overtraining you can guarantee a future where you'll wonder if some of your colleagues bought their medical diplomas from a Pakistani mill.

 

[DrfluffyMD]

Automatic contouring may happen at some point, but probably only for straightforward cases. It's going to hurt radiology first before it hurts radiation oncology. A typical scenario, where a machine can probably perform better than a radiologist would be lung nodule evaluation in chest CTs. A machine can be "fed" with a lot of CTs and will at some point be able to detect, classify and compare nodules a lot faster and more reliable than a radiologist. Automatic volumetric comparisons of repeat CTs of the same patient should also be more precise than a radiologist's judgement.

It’s not helpful to state one field is safer than the other due to the unpredictability of AI, but those who states that DR will be easier to automate than Radonc is just plain wrong.

A specialty that predominantly uses ONE theraputic medicine (photon) by applying it in a more or less predictable way governed by guidelines are far more of a low hanging fruit than automaton of diagnostic radiology. Neither are far from reality though.

Watson oncology is already reality. However no actual machine learning based radiology reading software is on the market despite years of hype.

 

[RickyScott]

with respect to computers/AI. Our job wont be taken over by robots, but computers/AI can simplify/streamline/dumb down radiation so we need less radoncs. Are we already at the point where a urologist (or well trained monkey) could quickly learn and deliver prostate radiation? breast surgeon?

 

[RickyScott]

with respect to rural vs cities: again, the preaching needs to stop. Rural populations are obviously deserving of quality health care.
The issue here is choice. Many medstudents are likely reluctant to enter a specialty where they have a low chance of working in a large metropolitan area (assuming thats important to them) or selecting a certain geography. Obviously, if a student (and spouse) has no geographic preferences, by all means...

 

[scarbrtj]

I just visited a major costal city, typically cited as one of the most "desirable" cities to live in and witnessed a homeless man masturbating on the sidewalk, human feces in alleyways, severe overcrowding, unreasonable costs of living, and general societal devolution.

San Francisco?

 

[maruchan]

With AI you need 1) an easy way to represent the inputs, and 2) a well-defined "ground truth" to feed to the machine learning algorithm

So IMO:
- Detecting a lung nodule = good application of ML. Input is just the CT scan alone and has a well-defined ground truth.
- Drawing the optic chiasm = good application of ML. Input is just the CT scan alone and ground truth is pretty well-defined as well.
- Making decisions on something like sacrificing PTV coverage or max hotspot to accept = bad application. Messy inputs with many variations including patient preference, physician preference, comorbidities, anatomy, etc etc. No clear ground truth, 4 different rad oncs will give you 4 different answers.

If AI reaches the point where it can make those more complicated decisions, then all of medicine will look completely different.

As far as breast surgeons doing radiation...I mean you can teach a pgy2 to do breast radiation reasonably in 2-3 months already. You could teach a breast surgeon in less time. The main barriers are sociolegal/political more than technical difficulty I think. There are more immediate threats to the job market than AI overall for sure.

 

[psy3]

With AI you need 1) an easy way to represent the inputs, and 2) a well-defined "ground truth" to feed to the machine learning algorithm

So IMO:
- Detecting a lung nodule = good application of ML. Input is just the CT scan alone and has a well-defined ground truth.
- Drawing the optic chiasm = good application of ML. Input is just the CT scan alone and ground truth is pretty well-defined as well.
- Making decisions on something like sacrificing PTV coverage or max hotspot to accept = bad application. Messy inputs with many variations including patient preference, physician preference, comorbidities, anatomy, etc etc. No clear ground truth, 4 different rad oncs will give you 4 different answers.

If AI reaches the point where it can make those more complicated decisions, then all of medicine will look completely different.

As far as breast surgeons doing radiation...I mean you can teach a pgy2 to do breast radiation reasonably in 2-3 months already. You could teach a breast surgeon in less time. The main barriers are sociolegal/political more than technical difficulty I think. There are more immediate threats to the job market than AI overall for sure.

I appreciate your insight on how AI can be integrated into clinical medicine and limitation there of. I wish I had better understanding of specific versions of AI to counter-argue. I will be interested hear what you would think if you had an opportunity to look into Google's Deepmind AI, AlphaGo (maybe you already have, then what I will say next would be a moot point).

IMO, what Watson achieved in chess and AlphaGo did in Go probably surpasses what does into contouring and evaluating plan. First, in-terms of messy variables in contouring, Go has 19x19 grid where you can put your stones, you go through about 200 turns, and there are 10^170 different ways the game can turn out. And every turn, the player (human or AI) needs to assess risk and benefit of the next move, and often foresee what would haven 100 turns down the road. Given the game is won or lost by how much territory a player secured, each move involves assessing the territory pattern at macro (the whole field) and micro (the immediate area where the next stone will go) level, which is in tune with decisions we make everyday (what to cover, what to avoid, how much to cover or avoid). And even though individual radiation oncologists may come up with various iterations or preferences, the ultimate outcome that matters are easily quantifiable, right? Coverage, in-field/out-of-field failure, toxicity in individual OAR.

Anyway, back to AlphaGo, that AI in last <5 years or so, got created, studied all previously played Go games in major championship levels with multi-million dollar prize money, then solidly beat multiple world best players. I've heard the second generation of this AI now and beat the first generation (the one that beat best of human mind) with its algorithm of just understanding the principle of gain, but not studying previously played games. So, the name of AI now is apparently changed to Alpha, as people in Deepmind/Google solidly believe that they conquered Go. It is my bias that that game of Go is infinitely more complex and hard of what I do even in the worst mess of head and neck plan. So, I personally believe that AI will be capable of taking over the contouring/planning part before my profession career is over.

But, like you said at the end of your post, we will have to see how the social and political barrier in incorporating AI in clinical cancer care. I fear, though, that if such capable AI tool is available to our gate keepers, surgeons, medoncs who actually see patients in person and they would tell the patients that they can do it with this cutting-edge programs and machines, would patients even bother to know about existence of radiation oncologists? We've all experienced the situation where referring docs already tell patients "you don't need radiation," and patients just listen and move on.

 

[Dare2Dream]

this really is one of the most important things in the whole thread from oldking above

"I know somebody who has wanted to move to Charlotte, NC, which is actually a large and booming city recently, but there hasn't been a job there in years. He has actually applied to that job mentioned above in Boone, NC, which is a satellite 1.5-2 hours outside of Charlotte (and the closest airport) in a college town (Applechia State which appears to actually have major division I sports, a Home Deport, Barnes and Nobles, etc so not that rural). Apparently nothing else in the entire state of NC, which I was surprised to just look up and discover somehow has a population over 10 million people, other than in the extreme East, which is far away, or apparently at UNC but you have to be the chief at Harvard to get a job like that!"

Let this sink in med students bc this is the reality. I know someone from top 5 program that needed to move to alabama bc his wife would be working there, he cant find anything within 2 hours distance so right now he isnt working. Let that sink in. You essentially have to be geography agnostic 100% for rad onc.

This is an important point. It isn't just that large coastal cities are saturated but that there is diminishing choice about location on a much greater level. It has been mentioned by a few posters that rural populations deserve local quality care which no one disagrees with yet no one stating that has also stated that they have themselves voluntarily made that choice. It is only a matter of time that the greater supply starts affecting those in stable metropolitan areas as Medonc groups or hospitals start bringing in cheaper in house Radoncs.

 

[deleted18755]

This is probably the best summary but I think it should be modified to a bit to:

"If you are flexible and okay living (ANYPLACE) in the Midwest or South, then you'll be more than happy (AT LEAST FOR THE NOW).

I know somebody who has wanted to move to Charlotte, NC, which is actually a large and booming city recently, but there hasn't been a job there in years. He has actually applied to that job mentioned above in Boone, NC, which is a satellite 1.5-2 hours outside of Charlotte (and the closest airport) in a college town (Applechia State which appears to actually have major division I sports, a Home Deport, Barnes and Nobles, etc so not that rural). Apparently nothing else in the entire state of NC, which I was surprised to just look up and discover somehow has a population over 10 million people, other than in the extreme East, which is far away, or apparently at UNC but you have to be the chief at Harvard to get a job like that!

UPDATE: I was a little concerned that maybe I was posting second hand information inaccurately so I emailed my friends the other day.

Still no luck in Alabama (and from another post above it sounds like there is at least one other person desperate for a job in Alabama).

Interestingly enough my friend who is trying to get to NC had interviewed with a well established private practice in a medium sized but booming city called Greensboro to staff a satellite 45 minutes away in dead/dying old mill town called Eden, NC. He is not interested in academics and actually doesn’t mind living in and caring for people in rural economically depressed areas so thought it would be a “win/win” since the others in the group didn’t seem to like going there. Long story short UNC acquired the hospital and everybody thought they would honor the private practice’s service agreement but they terminated it and they are now posting on the career site for a non-tenure track, “fixed term” position paying $230,000-$300,000!

Clinical Faculty

I guess they know they can get some desperate resident to work in a rural town 2 hours away from UNC's actual campus with no long-term contract cheaper than they can staff the place with the private group that has been there forever.

I honestly think the saddest part is that the UNC physician will dread working there, probably commute an hour and curse the town, and leave as soon as he can while my buddy would have lived right in the middle of the town and given back to the community outside of work while likely staying there until the day he retired.

At least UNC saved a few bucks (yeah right $600,000 ... $200,000-$250,000 is already the norm!)

 

[medgator]

UPDATE: I was a little concerned that maybe I was posting second hand information inaccurately so I emailed my friends the other day.

Still no luck in Alabama (and from another post above it sounds like there is at least one other person desperate for a job in Alabama).

Interestingly enough my friend who is trying to get to NC had interviewed with a well established private practice in a medium sized but booming city called Greensboro to staff a satellite 45 minutes away in dead/dying old mill town called Eden, NC. He is not interested in academics and actually doesn’t mind living in and caring for people in rural economically depressed areas so thought it would be a “win/win” since the others in the group didn’t seem to like going there. Long story short UNC acquired the hospital and everybody thought they would honor the private practice’s service agreement but they terminated it and they are now posting on the career site for a non-tenure track, “fixed term” position paying $230,000-$300,000!

Clinical Faculty

I guess they know they can get some desperate resident to work in a rural town 2 hours away from UNC's actual campus with no long-term contract cheaper than they can staff the place with the private group that has been there forever.

I honestly think the saddest part is that the UNC physician will dread working there, probably commute an hour and curse the town, and leave as soon as he can while my buddy would have lived right in the middle of the town and given back to the community outside of work while likely staying there until the day he retired.

At least UNC saved a few bucks (yeah right $600,000 ... $200,000-$250,000 is already the norm!)

Personally, I'd just as soon take a job in a rural area in the Midwest and make $400-500K+ given that specific choice

 

[deleted18755]

Personally, I'd just as soon take a job in a rural area in the Midwest and make $400-500K+ given that specific choice

From what I understand it was very easy to find a job in medium sized cities in the south for $450,000 starting and $600,000-$750,00 after partner as recently as 5-10 years ago so if jobs in the rural south (Eden, North Carolina - Wikipedia
population Eden, NC is literally 15,000 and it is 45 minutes from Greensboro, which looks like a nice medium sized town but most people wouldn't consider to be a booming metropolis) are being posted now for $230,000-$300,000 I'm sure it's only a matter of time before those physicians making $400,000-$500,000 in the MidWest will be pushed out and replaced by desperate new graduates who will be willing to work for less than half the salary.

I honestly didn't think it could happen this fast but here is at least one clear example...

 

[Haybrant]

As someone strongly associated with the ai/cds field you are far off what will happen in this regard

Decision making will be greatly added / substituted by machines in the future. And this goes for all sectors in medicine, not only radiation oncology.

Medical oncology is also going to get a hit, when machine algorithms are going to tell you what chemo-/immunotherapy regime you are to use, after a complete sequence of the cancer genome has been performed and other already documented patient factors like comorbidity, age, etc are taken into account. Big data will help to guide the right decision.
The same pattern will also be visible in other disciplines too, like infectiology or pharmacology. It's already happening right now, actually. I remember having to look up what medication I was prescribing to a patient and if any potential interactions between the drugs were to be expected. Nowadays, we prescribe electronically and the system will tell me if I am prescribing two drugs that may influence each other in any way. A system like that pretty much eliminates something like 30% of the workload of a clinical pharmacologist. Those guys used to come to ward rounds once per week in the past and point out at interactions of prescribed drugs, they don't do that anymore.


Surgical disciplines are going to get less of a hit, since autonomous robotic surgery is not to be expected soon and with robotic surgery the workload (as in hours spent in the OR) for a surgeon actually increased.
And some specialities still perform quite delicate manual exams, which cannot be completely overtaken by machines, like opthalmology.

Automatic contouring may happen at some point, but probably only for straightforward cases. It's going to hurt radiology first before it hurts radiation oncology. A typical scenario, where a machine can probably perform better than a radiologist would be lung nodule evaluation in chest CTs. A machine can be "fed" with a lot of CTs and will at some point be able to detect, classify and compare nodules a lot faster and more reliable than a radiologist. Automatic volumetric comparisons of repeat CTs of the same patient should also be more precise than a radiologist's judgement.

 

[DrfluffyMD]

From what I understand it was very easy to find a job in medium sized cities in the south for $450,000 starting and $600,000-$750,00 after partner as recently as 5-10 years ago so if jobs in the rural south (Eden, North Carolina - Wikipedia
population Eden, NC is literally 15,000 and it is 45 minutes from Greensboro, which looks like a nice medium sized town but most people wouldn't consider to be a booming metropolis) are being posted now for $230,000-$300,000 I'm sure it's only a matter of time before those physicians making $400,000-$500,000 in the MidWest will be pushed out and replaced by desperate new graduates who will be willing to work for less than half the salary.

I honestly didn't think it could happen this fast but here is at least one clear example...

So how do radoncs get paid? Where are the pro fees generated from? Do those 600-700k salaries include technical components which new grads are less and less likely to get?

 

[Reaganite]

From what I understand it was very easy to find a job in medium sized cities in the south for $450,000 starting and $600,000-$750,00 after partner as recently as 5-10 years ago so if jobs in the rural south (Eden, North Carolina - Wikipedia
population Eden, NC is literally 15,000 and it is 45 minutes from Greensboro, which looks like a nice medium sized town but most people wouldn't consider to be a booming metropolis) are being posted now for $230,000-$300,000 I'm sure it's only a matter of time before those physicians making $400,000-$500,000 in the MidWest will be pushed out and replaced by desperate new graduates who will be willing to work for less than half the salary.

I honestly didn't think it could happen this fast but here is at least one clear example...

I still remember the years when so many on this board bought into ASTRO's attack on urorads...private practice urologists who actually took care of their rad oncs with good pay, easy work, and in many cases desirable geography. Now we're left with the academic version of urorads with lower pay and hospital admins and chairmen who don't give a crap about our well-being. Instead of bitching, support and encourage local private referring docs to get together and open up more integrated freestanding centers where you have a cut and everyone can share in the benefits of the technical component. I'd be happy to help...

 

[medgator]

I still remember the years when so many on this board bought into ASTRO's attack on urorads...private practice urologists who actually took care of their rad oncs with good pay, easy work, and in many cases desirable geography. Now we're left with the academic version of urorads with lower pay and hospital admins and chairmen who don't give a crap about our well-being. Instead of bitching, support and encourage local private referring docs to get together and open up more integrated freestanding centers where you have a cut and everyone can share in the benefits of the technical component. I'd be happy to help...

This!

Freestanding centers provide the needed competition and lower cost alternative to the entrenched hospital/academic centers. Insurance companies will be interested in having the lower-cost freestanding alternative in their preferred network in some cases.

 

[DebtRising]

Very simple. Over half the field responded to ASTRO survey that oversupply biggest problem. ASTRO response 'shove off'.

Organization with that integrity will never inform medical students about realistic employment opportunity. Field is in crisis if half respond that there are too many. Hypofractionation make worse, even though ASTRO promote this even in absence of good evidence [love ASTRO high risk prostate guidelines].

When I interviewed looked at jobs in some 'rural' areas. At one point growing up family had horses in yard next door. It not because we live next to billionaire horse ranch, we lived in country. Lot of 'rural' jobs not pay anywhere close what you cite. No SBRT, SRS, brachytherapy capability. Not credentialed to enroll clinical trials. Often taken over by Big University X whose 'vision' seems to be grind new grad after new grad into ground so department can pocket technical. Salary, so no incentive of private group who was forced out.

Also, sick of myth of 'best jobs not advertised'. No. No university now is 'not advertising' tenure track or faculty appointments, least not on this large scale some claim. Certain places will find fit in one of own, great. But certainly for tenure track, they advertised, 95%+. Private practices like alumni network and some will not advertise, so definitely pay to network. But they not sitting on awesome job just waiting. They have network of friends and colleague feeding interested resident and often have multiple people interested. The job board is the majority of jobs. During interview time found 3 jobs by 'networking' that not posted - in my case all poorly executed, defined, and not position I accepted. People genuine and respond to genuine interest, but no employer served best by not post a job.

As for recent AI turn, simple trick is advocate for and manage patients. Explain the future, fight at tumor board, show up bedside when admit, follow results. In last 2 weeks approve 3-4 plans that would be 'rejected' by AI simply due to dose constraints. Is 10% grade 3 toxicity worth achieve dose associated with X% better control? Don't know, but patient can inform risk they want to take fighting deadly disease and doctor should responsibly describe risk and then make decision. So much institution variability in SRS, SBRT, brachy, what is even right answer or technique? Patients not text book - far more variables than CT set. Also changed management of 2 patients due to disagree, respect, with other physician the supported by subsequent biopsy. Be oncologist, that is not replaceable. Rad not cook book, not even close. Some regimens defined, but that changing rather quickly.

 

[maruchan]

I appreciate your insight on how AI can be integrated into clinical medicine and limitation there of. I wish I had better understanding of specific versions of AI to counter-argue. I will be interested hear what you would think if you had an opportunity to look into Google's Deepmind AI, AlphaGo (maybe you already have, then what I will say next would be a moot point).

Playing Go is a far easier task for AI for a simple reason: you can simulate games and outcomes are deterministic. The most recent iteration of AlphaGo works through a clever combination of machine learning and reinforcement learning, where the AI plays itself many times. You can't simulate outcomes of cancer patients, otherwise we wouldn't need RCTs. The outcomes like infield recurrence, rates of toxicity may be "quantifiable" but we have no way to create precise estimations of these numbers for a given plan.

We're not even close to AI being useful for decision support. Current tech is good at things like visual pattern recognition but can't do any "critical thinking" whatsoever. The most AI can help with is saving you time checking the NCCN guidelines (and it's not like doctors spend hours per day checking the NCCN guidelines), and maybe help some rogue physicians adhere to them better. Something like watson oncology is near worthless in its current form. On the other hand, auto contouring that actually works is quite close, but again IMO will have minimal effect on rad onc employment.

Honestly though this discussion is best saved for another thread. Oldking, haybrant, and others hit the nail on the head with the point below and that should be the focus of this thread.

this really is one of the most important things in the whole thread from oldking above

"I know somebody who has wanted to move to Charlotte, NC, which is actually a large and booming city recently, but there hasn't been a job there in years. He has actually applied to that job mentioned above in Boone, NC, which is a satellite 1.5-2 hours outside of Charlotte (and the closest airport) in a college town (Applechia State which appears to actually have major division I sports, a Home Deport, Barnes and Nobles, etc so not that rural). Apparently nothing else in the entire state of NC, which I was surprised to just look up and discover somehow has a population over 10 million people, other than in the extreme East, which is far away, or apparently at UNC but you have to be the chief at Harvard to get a job like that!"

 

[Haybrant]

Agree this is not the thread for AI. I am among the leaders in the country in what AI can do in medicine and Im telling you it will not have an impact on rad onc in a meaningful way for many many years, for certain it is not impacting the job market in any way right now.

I just spoke with someone else at a top 5 program, poor guy has to go to rural Connecticut next year for a job as his wife teaches at a college there and she cant find anything anywhere else that would be reasonable for them (you think rad onc is hard, try academic liberal arts jobs). 2 kids and family is about 1.5 hours away. He cant find anything within 2.5-3 hours of their house, and the offer he did get they know his situation and offered him 180k starting. They are in serious trouble w job and interpersonally, he thinks he will have to locums at whatever intervals he can find it. Med Students, listen up, you dont want to be in this type of situation. Stay Away from Rad Onc, dont even think about it at this juncture, our leadership has failed us we all know about it

 

[medgator]

Med Students, listen up, you dont want to be in this type of situation. Stay Away from Rad Onc, dont even think about it at this juncture

Honestly that isn't really fair to someone if they don't care where they end up. There are still decent jobs from a financial standpoint, even now, if you are willing to go to the right locale away from the coasts.

If given the choice of doing rad onc vs another specialty or PC in a rural setting, I'd do rad onc all the way.

 

[Neuronix]

Where I trained in a semi-desirable location they had high faculty turnover so there were jobs available every year. As a trainee the faculty were constantly telling me how miserable they were and how they were looking for new jobs. Despite this, the departmental leadership offered a fellowship to all of the graduating residents who they liked. If the resident got a real offer somewhere else in academics, only then would they give an offer more than a fellowship. I guarantee you that as soon as the job market tightens to the point where no academic jobs are available, they'll have plenty of fellows. I was involved at the administrative level there as a resident. One of the admin staff involved with the residency program was also involved with faculty recruiting. She told me that rad onc had the lowest priority because they knew the job market is bad. Contracts were non-negotiable. They had no problem filling the jobs so why bother offering anything?

This was a few years ago now. We're still graduating way too many residents. Things just get worse every year. I agree with the other posters. Things aren't horrible *for now*, but they're getting bad. The next few years will be even worse with the 200 graduating residents per year and non-expanding job market.

You guys talk about "the south" being okay. You already heard about Alabama. I have another friend who desired strongly to be anywhere within another southern state to be close to family with their small children. They're in the next state over in a small city several hours away because nothing was available in that particular state.

As for academic satellite expansion jobs--there are plenty of these. This is basically the only part of our job market that is expanding. They put you 100% clinical in a satellite for lousy pay because you have no other option. Promotion pathway/pay raise still has to do with academic productivity. How are you supposed to be productive academically out at a satellite working 100% clinical with no resources? The problem is though that the private groups also are offering poor salaries nowadays. $200-$300k are common even in undesirable parts of the south now from both academics and private practice. It's a race to the bottom.

 

[medgator]

? The problem is though that the private groups also are offering poor salaries nowadays. $200-$300k are common even in undesirable parts of the south now from both academics and private practice. It's a race to the bottom.

If you check the arro surveys, $250-275K was the average starting salary for a partnership track position a decade ago. I guess not a lot has changed