Interviewer: Dr. Levatter, thanks for taking more time from your busy day to chat with us. We got a lot of positive responses from your first interview. Readers seemed to think you were more willing to speak freely and candidly about both the pluses and minuses of your specialty.
Ross Levatter, MD: Well, I’m glad at least some of your audience enjoyed what I had to say. I’ve known a lot of very smart and clever radiologists in my career. I think though what sets apart aspects of what I’ve done over the years from that of other radiologists is my dedication to problem-solving, to answering the clinical question. That’s not as easy as your readers may think.
I: What do you mean?
RL: Let me give you an example with a fascinating case from back in the 1990s. A woman in her 30s presented with a several month history of intense back pain. Prior chest and rib radiographs were normal, as was chest CT. But a bone scan (a nuclear medicine test where a radioactive tracer is deposited throughout the skeletal system) showed a focal area of increased uptake (“hot spot”) in the back of one rib. Despite repeat rib detail imaging, nothing could be seen. That’s when her surgeon got me involved. It occurred to me we needed to combine the bone scan’s sensitivity with a highly specific exam like CT. Back at that time, routine CT slices were 8 mm in thickness. That is, each image represented an 8 mm thick block of tissue. So if, for example, the lesion you were looking for was 3 mm, and only half of it was on any given image, you could miss it because of insufficient resolution. You could DO thinner slices, of course, but given the scanners of the day, you couldn’t do them through the ENTIRE chest, like we can now. So the problem was “Where do we take the thin slices to help better see the lesion?”. Here’s what I did: I repeated the bone scan. I then used a small handheld radioactive probe to locate the skin just above the hot spot. With that point marked, I walked the patient over to the CT scan and did specialized high-resolution CT just over that area. This allowed me to demonstrate a focal lytic lesion subtly expanding a small portion of one rib. I diagnosed an osteoblastoma, which was the most common lesion to cause pain and look like that, and it was confirmed at surgery. I subsequently used this technique, combining the sensitivity of nuclear medicine with the specificity of CT, over the next year to diagnose several subtle lesions.
Now why do I say that’s anything different than what most radiologists would do? Because no one else in my group was interested in this technique despite its proven effectiveness. The main reason, an understandable one, was that the technique was time-consuming. Some of the other guys in the group tried to do this by just telling the technologists in CT and Nuclear Medicine, “do that thing you do for Dr. Levatter,” but sadly it was a little more complicated, more radiolgist-involved than that. And the way medical care is paid for today, if there isn’t a government-generated medical code for what you did, you can’t get paid for it. So I was taking time that wasn’t being reimbursed, as there was no coding that covered the extra time it took to coordinate the two studies. I didn’t mind doing this, because I liked the results I got, but I can’t fault other doctors for thinking it would be nice to be paid for their helpful efforts. The irony is that today, 20 years later, the technique I developed is very similar to combined PET/CT exams that cancer patients routinely get. But in the early 1990s, no one else was doing this.
I: Another fascinating case, Dr. Levatter. I guess this is another example of what you mentioned last time about how over time technology catches up with clever diagnoses, so that everyone can do now what only a few could do before.
RL: You’re right about that. Here’s another example: In the early 1990s, while I was in on a weekend reading ER films, I came across a case of a patient in a motor vehicle accident the prior day. I was looking at her chest x-ray, which at that point was about 15 hours old. It had initially been read out (as was customary at the time) by the ER physician, who thought it was normal. I called him as soon as I looked at it, telling him the mediastinum (the part of the chest between the two lungs) was subtly widened, and in the context of a car crash we had to rule out a traumatic aortic transaction, that is, a tear of the major artery of the chest. This is not something that should wait 15 hours. The ED physician immediately called the patient, who had been released—she was actually feeling fairly well—and demanded she return immediately. I did a chest CT on her as soon as she hit the door. Now, today, with fancy CT techniques--very thin slices, computer reconstruction, etc.--CT is the modality of choice at making this diagnosis. But then, 20 years ago, articles suggesting CT could ever make this diagnosis were just appearing in the literature and “everyone knew” the diagnosis was always made/confirmed with angiography, which involves threading a catheter from the groin artery into the aorta and injecting dye, an invasive procedure done under OR-like conditions. But I prescribed the CT using a special technique which allowed me to create computer images that mimicked what you’d see on angio. No radiologist at my hospital had ever done this before, it seems. When the cardiothoracic surgeon came in to review the images with me, he said, “I’ve never done this surgery before without an angiogram.” I replied, “I can have the angiographer here within the hour if you think you need it.” He responded, “No. These CT images are completely convincing. I’ve just never seen anything like it before.” The patient was saved, very lucky, since 95% of aortic transactions that far out die. But today, everyone makes this diagnosis by CT. With the new CT scanners, it’s become much easier.
I: Are most radiology diagnoses just a function of technology?
RL: No. Sometimes you make diagnoses now just like you did 20 years ago, using basic medical and imaging knowledge. For example, a few years ago I reviewed an abdominal CT on a middle-aged woman who had a large mass in the right kidney and a similar large mass in the left kidney. These were unchanged from a similar study of one year earlier. At that time, both had been called large simple kidney cysts. While stability over one year was certainly reassuring, the problem was the diagnosis of simple cysts was that, though they were low in density, they were not low enough in density. CT allows us to measure, quantitatively, precisely, the density of a structure, in CT units. A simple cyst should have a density of 0-20 CT units. These masses I was looking at were measuring 60-80 units, and therefore these lesions were really solid, not fluid. Although solid masses in both kidneys would be much less common than cysts in both kidneys, the CT numbers made me raise concern about bilateral renal cancers, which is what was found at surgery. Luckily, they were very low-grade, which is to say, slowly growing, and the patient did well.
I: I recall one of your subspecialty interests is in imaging the gut, or GI tract.
RL: Yes. Let me share some of my favorite cases in that area.
I recall the case of a sick teenage girl who came to the Emergency Department with abdominal pain primarily in her right lower quadrant. The CT showed multiple abnormalities including an abnormal appendix. The initial radiologist called it acute appendicitis and urged urgent surgery, but clinically that didn’t seem to fit, so I was asked to give a second opinion. On reviewing the CT, I told them that multiple loops of small bowel were abnormally thickened and abnormally enhancing, including but not limited to the appendix, and that especially in a teenager the most likely diagnosis was Crohn disease. Crohn disease is an inflammatory bowel disease that lasts a life-time, getting better and worse over time--waxing and waning, we say. It should be treated medically if at all possible and surgery is avoided unless absolutely necessary. So the subsequently proven diagnosis of Crohn disease led the doctors taking care of her to change her patient management, and avoid surgery.
I: So radiologists can have a big impact on patient care.
RL: Oh, yes. Absolutely. Here are two more examples of exactly that:
First, there was a patient with right upper abdominal pain who underwent a CT. The original radiologist noted a stone in the gallbladder and abnormal thickening of the gallbladder wall, and so correctly diagnosed acute cholecystitis. He also claimed there was acute appendicitis. This was confusing to the surgical team as there was no clinical evidence of that diagnosis, so they showed it to me. In my second opinion, I pointed out several findings that had been missed. First, some abdominal blood vessels were enlarged, implying the patient had portal venous hypertension, usually a result of increased resistance to blood flow through the liver. This is most commonly associated with cirrhosis, though the liver didn’t look particularly cirrhotic. My observation led to a work up for cirrhosis which was positive and unsuspected. Second, there was an enhancing mass, about an inch in diameter, just in front of the aorta at the L4 (4th lumbar spinal) level, completely overlooked on the initial reading. The finding, though rarely seen, was typical of an extra-adrenal pheochromocytoma. What’s that? Well pheochromocytomas are hormonally active tumors found mostly in the adrenal glands, which sit on top of the kidneys. Only 10% of them are found outside the adrenal gland, but just in front of the aorta at L4 is the most likely spot for them. These tumors, which may be malignant, can cause a huge spike in blood pressure during general anesthesia. It could have been catastrophic to operate to remove the gallbladder without guarding against a large blood pressure spike. Therefore the anesthesiologist needed to be warned before the surgeon removed the gallbladder. Finally, the alleged acute appendicitis was really an appendiceal mucocele, a tumor that dilates the appendix due to filling it with a gelatinous material. These can also be malignant, though they are usually benign. They are removed surgically in all cases, but the surgical approach is very different from doing a standard appendix removal for appendicitis, because if the appendix ruptures and the gelatinous material spreads through the abdomen, life-long significant complications can arise, so, again, the surgeon needs to be forewarned. Surgery proved me right about the pheochromocytoma and the appendix.
Or consider this 2006 case: a 35 year old patient had a CT for nausea and vomiting, read out by the first radiologist as normal except for a small amount of fluid in the region of the gallbladder fossa, or cavity, where clips showed prior gallbladder removal. The fluid was attributed to recent surgery, which would have made sense except for the fact that when I questioned the patient she told me her gallbladder had been removed years earlier. The surgeons wanted me to do a special study of the small bowel, not done most places, in which I had developed some expertise. It’s called an enteroclysis, and involves threading a catheter from the patient’s nose into her small bowel and performing a detailed study of the gut. But in preparing for this I reviewed the CT. On my review, I decided the correct diagnosis was on the CT but had been missed, so to confirm it I suggested a different exam, called a hypotonic duodenogram. The initial radiologist hadn’t noticed the stomach and first part of the duodenum were dilated but the second portion of the duodenum was narrowed, and there was abnormal pancreatic tissue surrounding the duodenum. Normally the head of the pancreas sits next to the duodenum, but it never normally surrounds the duodenum.
I: So what did you do, a hypotonic duo-what?
RL: I did a specialized barium study of the duodenum, a hypotonic duodenogram. It’s hypotonic because you begin by giving a small amount of medicine to relax the gut, make it hypotonic. Then you manipulate barium and air into the duodenum and take pictures. As I suspected they would, the pictures showed the portion of the duodenum next to the pancreatic head to not have its normal cylindrical tube shape; instead it looked like an hourglass, narrowed in the middle. That was because the head of the pancreas was abnormally wrapping around it. This is why the patient was throwing up. It’s called an annular pancreas. The fascinating thing is that while the lesion is congenital--people are born with it--surprisingly, it often doesn’t present until adulthood. This diagnosis led to the proper surgical correction.
I: So was this a bad miss by that other radiologist?
RL: Well, it was bad in the sense it didn’t help the patient. But it wasn’t bad in the sense that most doctors would pick up the abnormality. It was very subtle on the CT. But as I told you, the GI tract is a specialty interest of mine, so I always study it very carefully. In addition, I’d guess as someone interested in the GI tract I’ve probably done many more hypotonic duodenograms than most radiologists. It’s not often done, but this was a perfect example of how it can be helpful.
I: I understand you need to get back to work. These cases are fascinating. Can you meet with us again.
RL: Happy to. I hope your readers are enjoying this as much as I am.
