271 karma · joined October 15, 2016
Idk, maybe for whatever reason right now, as others mentioned, the macro environment or maybe your micro environment is incentivizing this behavior but it won’t work out in the long run.
Just gotta do you, and do good, OP.
Edited for clarity
Depending on the protocol, CT scans usually take anywhere from 5 seconds to 10 minutes (in the longer scans most of that is waiting after contrast injection).
MRI studies take a long time to acquire data, usually at least 20 minutes for very short/limited studies. More involved protocols can take hours. Patient throughput is slower.
It’s cool, it’s a nice study. The article definitely over hypes the “common cause of hypertension”. These lesions are rare. Most people with high blood pressure have it because of metabolic disease/chronic stress/lifestyle, and it can be managed with changes to diet exercise and meds.
It is true that doing adrenal vein sampling is hard and not always successful. Having a protocol that can non-invasively detect these lesions may save some people a procedure.
1. The confidence intervals for the risk differences are wide, and some categories include negative numbers. This means that perhaps the vaccines were protective (?) against these other outcomes, but overall the trend based on this analysis is the vaccines were associated with increased risk of these other bad outcomes.
2. Capturing and reporting SAEs in the study designs are sensitive but not specific (i.e. they try to capture everything and aren't particularly concerned with verifying the cause of the SAE). For example, in the Pfizer study, they captured data related to any SAE from the start of consent to 6 months after the last dose of trial vaccine [1]. That is a long time for typical medical problems to crop up. I would like to know a comparison between the intervention group and the estimated rates of such illnesses for an age-matched population.
3. In this study they looked at all SAEs as compared to any SAE (which is what the FDA did). I don't want to discredit they study in any way and it is a valid analysis. However, it is important to keep in mind that when someone becomes ill, injury to one organ system can cause downstream injury to a second (for example a heart attack reduces cardiac output, which causes kidney injury). One way to explain the findings here is that the vaccine triggered more secondary, unintended, bad outcomes. Another way to explain the data is that a small number of people became very ill in the peri-vaccination period and had a lot of bad stuff happen. Also keep in mind from point 2 that the time period for collecting SAEs was from consent to 6 months after the last dose of vaccine.
The data presented here is interesting, and it's possible that the vaccines could be more harmful than initially thought. Overall, I am skeptical. The numbers here are still very low compared to the total number of people in the trial. If patient level data is released and there is more meat to the findings here it might change how I think about vaccination, but at this point I wouldn't change.
[1] https://www.nejm.org/doi/suppl/10.1056/NEJMoa2034577/suppl_f...
The main part is that our success is highly situational. I think this somewhat ties in with luck - working hard in the right situation pays off.
Results are not impressive until they are :)
It's certainly not a solved problem, and it's easy to have a pessimistic view now but I'm generally bullish on where things will be 10 years from now.
Not to mention imaging protocols are not standardized, and the imaging technology is also evolving so scans we do today may not be "correct" or standard in 5-10 years.
I am hopeful though. Few-shot learning and self supervision are very active questions right now and there are a lot of papers in the medical AI field that are getting published on these topics.
I'm personally interested in liver cancer, which does not have large, well-curated and shared database of cases.
Sharing data gets tricky, especially in the US. Labels I'm working with and creating (at least at this point) are for my own research, which won't be shared publicly any time soon.
It's not the case the more screening is always better. There are plenty of screening regimes that you've never heard about because the trial data didn't support it.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3762603/
What I found mind blowing when I first read these studies (particularly the second link) is that the positive-predictive value of the CT screen was only 3.8% for lung nodules that were 4mm or larger in size. Basically just means we find all kinds of lung abnormalities all the time that "could be cancer" but aren't.
Keep in mind, that the screening criteria was people over age 65 who had a 30+ pack-year smoking history (1 pack-year = smoking 1 pack of cigarettes per day in one year), and had to have smoked within the last 15 years.
I can also tell you from doing CT lung biopsies that it's not a procedure to take lightly. Lot of important structures in the chest.
It's our job (as radiologists and engineers) to shape that in a way that benefits patients.
I don't think that a human reader in the loop will be going away for a long time. Sure, they can be at parity, but does a combined read have better results? This is an active research question.
I think what will happen is that over time, human interpreters and AI systems will "co-evolve" in a sense, where people will pick up on where models are wrong and also learn how to use models to understand their own blind spots. These are also active research topics that are in their infancy.
EDIT: To be clear, I believe that this is a different group of authors from coolwulf. Pubmed link:
Does a patient take a negative exam as "I'm good!" and forego actually seeing a doctor, or having the exam read by a radiologist?
A positive result is also a challenging situation. I've talked to patients before after diagnosing them, and as you could expect most people are shocked, scared or they want to know what they should do next.
In breast radiology specifically (at least in the US) there is a well defined reporting lexicon and classification system. From what I can tell, the system does not use the lexicon or BI-RADS classification. https://www.acr.org/-/media/ACR/Files/RADS/BI-RADS/BIRADS-Re...
Breast cancer can be very subtle, and can be hard to figure out, especially if a patient has had surgery or other benign abnormalities. https://radiopaedia.org/cases/development-of-dcis?case_id=de...
Diagnostic and screening images also undergo rigorous QA for the entire system. Does the model have some kind of QA built in?
There is a just a small disclaimer at the top that says "we will not store your data on the server. please do not worry about privacy issues". Not exactly a formal or legal binding agreement.
In fact, you can't even connect over https. http://mammo.neuralrad.com:5300/upload
These are just a few things that come to mind off the top of my head. Like I said before, I do think that what coolwulf works on and is trying to do is good, and in the long run can help doctors better characterize findings and help patients. But like anything in health care, there are a lot of edge cases and side effects that you have to think about. The stakes are also very high.
I think in some ways making the model available online can be good, but in other ways could be harmful too. Very complicated topic.
恭喜coolwulf， 祝你继续成功。
2. Publish 2 papers
3. Go through all board prep material (one pass before 2023)
4. Eat less dairy (cut meat out last year)
[edit for formatting]
1) Are there ways you can keep going but change something to help? Going to see a therapist? Taking a short break? Talking about what's going on with your family? When you're stressed it's hard to remember all the support structures out there.
2) You need to ask yourself what is it worth to stay where you are (from a financial point of view). Is it worth a big payout in a few months to a year? Getting acquired is a good way to earn a big pile of FU money quick - and that will buy you all the time to relax and recoup that you need.
Do what you need to do to get better, and dont trash what youve worked hard to build! You can do it! Hope to see you post your success story a few months/years from now!
This is something I don't understand. I'm a medical student, and the amount of time I've seen wasted watching older physicians type notes is staggering. Just have them dictate the notes and hire a secretary. Perhaps for legal/liability reasons doctors need to write their own notes but it's just such a waste of time, and we pay a lot for it.
I think the part that struck me the most was his comments about time. I have diverse academic interests. I studied math and bio in undergrad. I love machine learning and software development (esp python). I lived in China to study the language for a year. All that gets sucked out of medical school though. We are expected to learn a ton of material in the first two years. Then in the second two years, we are basically working a full time job in the hospital/clinics while also studying. We are constantly evaluated. We are also expected to do research and publish papers. I've forgotten what a guilt-free day off feels like.