- Joined
- Jun 6, 2014
- Messages
- 8
- Reaction score
- 4
I am a current MS4 planning on applying to rad onc this year, and I truly love the field as it stands now, but there are some aspects of it that I was worried about and was hoping for some thoughts from this board. I understand that reimbursements will go down in rad onc as in all medical fields over the next 10-20 years, and I am fine with that. What worries me more is a fundamental change in oncologic care or radiation oncology and how that might affect radiation oncologists (in a similar vein to Dr. Zietman's article from 2008).
1. One example is automated contouring. Programs can already contour some normal structures, and with the breakneck speed of computer science these days, I wouldn't be surprised if all contouring of tumor and normal structure volumes, with appropriate margins, in the next 15 years is done automatically. This would mean that each rad onc can save the 20% of the time they spend on contouring and see more patients, which may mean you need 20% less rad oncs.
2. Another example is a "silver bullet" even for some cancers. Attendings I've talked to are usually dismissive of this, saying that there are only a few drugs that substantially prolong life by themselves. But that's SO FAR. 20 years ago we barely knew anything about cancer biology and now there is so much known that if just a few of these things lead to "cures" of common cancers (or at least obviate the need for radiation), it could be problematic for rad oncs. What if there is a targeted cure for prostate cancer or SCC of the H&N? And while it is true that all cancers are different, many of them exhibit some fundamental similarities that, if exploited, could lead to simultaneous cures for multiple cancers at once (e.g. inactivation of the p53 pathway). Radiation therapy will probably always be required for at least some cases, but society may need much fewer radiation oncologists.
3. Limited billing codes. This ties in to the above. Rad oncs can only irradiate. Due to the limited billing codes, if there is a targeted cure for one or some cancers making radiation less important for those tumors, rad oncs will not be easily able to shift their practice to alternative treatments that are not already being done by current rad oncs. I am hopeful that as radiosensitizers are developed, rad oncs will own them, but in my rotations, I saw that the med oncs still were the ones in charge of radiosensitization.
4. Streamlined clinical decision-making. I have seen how important having a rad onc in the room in multidisciplinary clinic is to make the decision about whether to use radiation, and if so, what type. But similar to my first point above, could there be a future where, based on the imaging/pathology characteristics of the tumor and staging done by the med onc, the optimal plan (both radiation and chemotherapy) is designed by a "clinical decision-making support tool" a la Watson? In this case, you would only need one oncologist (and this would likely be the med onc) to see the patient, talk to them, hold their hand, etc. and dosimetrists and physicists to actually deliver the radiation. You might need 1-2 rad oncs per hospital to tweak the treatment plans for special cases, but I could see a lot of the decision-making and contouring automated by advanced AI.
Am I overthinking all of this, or does any of this have the potential to fundamentally change the rad onc job market or how fun rad onc is in my lifetime?
1. One example is automated contouring. Programs can already contour some normal structures, and with the breakneck speed of computer science these days, I wouldn't be surprised if all contouring of tumor and normal structure volumes, with appropriate margins, in the next 15 years is done automatically. This would mean that each rad onc can save the 20% of the time they spend on contouring and see more patients, which may mean you need 20% less rad oncs.
2. Another example is a "silver bullet" even for some cancers. Attendings I've talked to are usually dismissive of this, saying that there are only a few drugs that substantially prolong life by themselves. But that's SO FAR. 20 years ago we barely knew anything about cancer biology and now there is so much known that if just a few of these things lead to "cures" of common cancers (or at least obviate the need for radiation), it could be problematic for rad oncs. What if there is a targeted cure for prostate cancer or SCC of the H&N? And while it is true that all cancers are different, many of them exhibit some fundamental similarities that, if exploited, could lead to simultaneous cures for multiple cancers at once (e.g. inactivation of the p53 pathway). Radiation therapy will probably always be required for at least some cases, but society may need much fewer radiation oncologists.
3. Limited billing codes. This ties in to the above. Rad oncs can only irradiate. Due to the limited billing codes, if there is a targeted cure for one or some cancers making radiation less important for those tumors, rad oncs will not be easily able to shift their practice to alternative treatments that are not already being done by current rad oncs. I am hopeful that as radiosensitizers are developed, rad oncs will own them, but in my rotations, I saw that the med oncs still were the ones in charge of radiosensitization.
4. Streamlined clinical decision-making. I have seen how important having a rad onc in the room in multidisciplinary clinic is to make the decision about whether to use radiation, and if so, what type. But similar to my first point above, could there be a future where, based on the imaging/pathology characteristics of the tumor and staging done by the med onc, the optimal plan (both radiation and chemotherapy) is designed by a "clinical decision-making support tool" a la Watson? In this case, you would only need one oncologist (and this would likely be the med onc) to see the patient, talk to them, hold their hand, etc. and dosimetrists and physicists to actually deliver the radiation. You might need 1-2 rad oncs per hospital to tweak the treatment plans for special cases, but I could see a lot of the decision-making and contouring automated by advanced AI.
Am I overthinking all of this, or does any of this have the potential to fundamentally change the rad onc job market or how fun rad onc is in my lifetime?