Week 7 - Aaron LaViolette

The last week of summer immersion has come, so this means the last blog post!

This week was mostly involved in shadowing Dr. Prince. While I was watching Dr. Prince read cases this week, I realized that I have come a long way in terms of how to read an image. When I first started watching images being read, I was really only certain of the modality of the image and the general anatomy (i.e. I could identify the organs). Now that I have reached the end of the immersion, I am now able to pick up that something is “right or wrong” with a particular organ. I am now also thinking differently. I am now asking myself questions like, is the lesion big or dark on T1?, T2?, does is show restriction on DWI?, was contrast given?, does it enhance with contrast? Indeed, this week Dr. Prince brought up a case on the computer and let me comment on what I saw. I used these questions to try to gain some insight. The image had a mass in the liver. I was able to see it was T2 bright, but not super bright like a cyst, and T1 dark. I was also able to see it was well circumscribed. The lesion was a hemangioma. When Dr. Prince told me what it was, it made sense based on the clues I was seeing.

Although I did not know all the vocabulary to describe what I was seeing necessarily and was not able to say what the lesion was, I was able to use clues to eliminate things the lesion wasn’t (such as a cyst). This is certainly an improvement over when I first started immersion. This made me realize that it is simple to train a person and given time they can learn. After all, in a short time of watching people, I was able to look for and determine clues as to what is going on, and if I was given a full rigorous training, I’m sure my results would be better. It also made me realize why it is simpler to train people instead of computers. Simply put people can learn things faster and with less examples, as compared to a deep learning algorithm. More importantly, people can do something a computer can’t: think. This reminds me that anytime computers are employed to help “find” a pathology, the radiologist should always be able to verify and override the computer’s answer if necessary.

Along these same lines, I also got to measure organ volumes for a few ADPKD cases and the T2* for a few liver iron studies (all verified by an MD of course). This week I was getting very close numbers to either Dr. Prince or Dr. Yin. This also emphasized the point that it is easy to train a person.

This week in the reading room Dr. Prince also made a couple of interesting comments about cancer. The first was that cancer is more likely among organs that are able to regenerate after trauma, such as the liver. This explained why many cases that where the patient had liver trauma (particular from hepatitis) in the past were being monitored. The idea that new growth has a better chance of becoming cancer is not particularly eye opening. However, it emphasized the importance of having a scan if you suspect any disease may be likely to occur, because early diagnosis is the best way to treat a disease. This is especially true with techniques such as MR, because there is no ionizing radiation.

The second point Dr. Prince made about cancer is was in the context of a skinny patient. On the MR scans we could see over time how the patient had lost weight, and just how skinny the patient was now, due the patient’s cancer. They are likely so skinny due to the cancer requiring so much energy. However, Dr. Prince mentioned how some people just simply eat more and are able to live for a quite a bit of time with the cancer, but how others simply lose the will to eat and die much sooner. This is an interesting point I did not realize, but it makes sense. It goes to show me the power of the mind. If they cannot overcome their negative thoughts and lose the will to eat, things turn bad quick.

The only trouble with MR, is if you have metal or magnets. Dr. Prince is often reading cases for breast reconstruction. In these cases, the patient wishes to have fat removed from them and placed where a breast use to be. Blood supply is important, and the surgeons want to know where the perforator vessels (vessels running into the fat) exist and how they interact with the muscle under the fat. Hence an MR scan is done to map the vessels. These patients many times have tissue expanders, which contain magnets.

This is a safety concern because the magnet in the expander will align with the magnet in the MRI, so the tissue expander will flip 90 degrees. However, Dr. Prince mentioned generally the 90 degree flip causes issues for the following reason. This is noticeable to the patient if they are laying on their back (because they can see it), but if they are laying on their stomach, they can’t see it only feel it. So, in his experience people tend to scream in the machine because they see something strange happening. In addition to this generally the higher the amount of fluid in the expander, the harder it is for the patients to notice the flip by feeling it happen.

However, sometimes these safety concerns as minimal compared to the risk associated with not being scanned, particular if the patient is worried about metastasis or new cancer. This week Dr. Prince had to clear a lady to enter the magnet with the expanders due to concern over metastasis to the brain. After the risks were explained to the lady she agreed to enter the scanner. The lady reported that she could feel the expanders flip 90 degrees when was near the magnet, and flip back when away from the magnet, but had no pain. Even with the flip, there was not much of an issue. Latter Dr. Prince commented that a lot of MR safety is based on extreme cases. This makes sense to me because like anything you design for the worst-case scenario. I learned that this design for the worst cases scenario principle is not different in the “design” of MR safety.

This week Dr. Prince also did some more research scanning, but this time with a phantom. A phantom is simply a large plastic tube containing fluid and other elements that should be able to be resolved on the scanner. Dr. Prince mentioned how this is used for periodic quality control. Because no one was in the scanner this time, Dr. Prince also used this as a teaching session to show a resident how to operate the scanner. Dr. Prince would suggest changing certain parameters to see what would happen to the image quality. This provided a great learning opportunity for me and really generated intuition on what happens to an image if a parameter is changed. I think hands on learning is particularly useful. Dr. Prince decided to go ahead and continue to train residents weekly as a sort of informal class in how to scan. I am disappointed I will not get to see more of these.