How I Do It- Imaging the Post-Treatment H&N Cancer Patient

Post-Treatment Cancer Imaging of the Head and Neck

I'm gonna talk about how I do post-treatment cancer imaging of the head and neck.

It's a massive topic. I can't really shoehorn it into 30 minutes. It's a very experiential thing. Obviously, you have to make a lot of mistakes in order to get better at imaging the post-treatment head and neck. And that's where my real expertise comes in. I've done so much of it that I've made, I can continue to make mistakes, but I learned from it, and I teach my mistakes and I share them.

So it's pretty difficult imaging. And it is not that we're just looking for the efficacy of therapy in a post-treatment cancer patient.

So head and neck cancer treatment causes considerable anatomic distortion that can unfortunately both mimic and mask tumor recurrence and may actually be coexistent with recurrent head and neck cancer.

In order to do this, the radiologist needs to be familiar with the disease process, and it's proclivities, what's the nature of, what's the behavior of that cancer? He needs to know when the treatment took place, what type of treatment, how that treatment might look in the face of successful therapy. So what would we expect to see as a side effect of that therapy? And if the tumor was effectively treated, what are the likely patterns of recurrence locations and appearances? And what are the complications of that therapy and their imaging appearance?

So really, the only way you can possibly do all that is to have a lot of data. So if you're sitting in the imaging center, and the patient was treated somewhere else, and they've had a complicated course, and now you don't know, you may not have previous imaging, the requisition may be incomplete, you don't have ready access to the treating physician. That's a very difficult place to read from. And so I'm sympathetic to anybody who doesn't have access to that kind of information, which includes other things.

I've already mentioned this, but what's going on with the patient right now? So on the day you're scanning the patient, what's going on? Does the patient have any new signs or symptoms? Or is there a new cranial neuropathy? Is there a new hoarseness? Anything like that? Why was the scan obtained? Was the scan obtained in a routine fashion? If it's three months after the surgery, that's my assumption, unless I hear otherwise. But if the scan was done two weeks after surgery, or four weeks after completion of radiotherapy, I'm suspicious. Even if their requisition doesn't say why they've done it, I'm suspicious that either they've made a mistake and scanned the patient too early, or there's some new sign or symptom that the guy's getting worried about, the clinician's getting worried about.

But these are things that I need to know. I also need to know where along the timeline, is it, okay, am I reading the first post-op scan or am I reading the third post chemoradiation scan? Where are we and do I have previous imaging? And what did the previous scans show? I mean, if I don't really know all these things, it's gonna be hard enough to read these scans, but if you don't have all this information, which really hinges upon improved technology.

So in my hospital, we have a really advanced electronic medical record. I have access to every iota unless the patient was seen that morning and hasn't, and the clinician hasn't transcribed his note, I know what's going on with that patient. I know their labs, their histology reports, their ocs, and I have access to all these things. And really the onus is on me. So I think that the electronic medical record changes have given us the opportunity as radiologists to be more like doctors to be more onco like oncologists. Alright? We're not just reading the film anymore, although that's the bottom line of what we're doing. But if we have access to this information, we have to use it, or we'll look bad and we'll make mistakes. And so I don't mean to preach, but that stuff's important.

Imaging Timing and Modality

Typically, I don't wanna scan somebody until at least eight or 10 weeks after therapy as a baseline. As far as the imaging modality, generally speaking, I want the same modality as they imaged the patient to stage the cancer. I don't want to go switching from CT to MRI, vice versa. And that makes my life more difficult. And unfortunately, we occasionally see some of that.

I would say that the more time you spend on the history, the less time you spend on looking at the scan, so that would be my advice. Spend more time on the history, figure out what's going on, what brought that patient to you that day, because otherwise, reading the scan is a little bit like opening up a book to chapter five. If you don't skim chapters one through four, chapter five is gonna really be tough. But if you skim chapters one through four, at least you have a fighting chance to understand what's going on.

And so I try to give a report that my history is stated. I don't just say head and neck cancer. The requisition may say head and neck cancer, which I hate, by the way. And I'll abuse the clinician who sent me such a requisition. But my requisition will say, advanced tongue-based carcinoma status post induction therapy followed, blah, blah, blah. And you can read that. But basically, I tried to say exactly where today's situation falls along the timeline of that patient's course.

Reasons for Imaging

Why are we imaging the patient? Sometimes to assess that a lesion has been completely resected in the case of surgery, to assess for complications of therapy, leaks and bleeds in the case of skull base or Cy Nasal surgery, to assess the efficacy of non-surgical therapy to exclude tumor recurrence. And this is a pretty big question. Why do we want to exclude tumor recurrence? In a lot of cases, the salvage therapies are not that palatable. Not in all cases, but is there a treatment for the, and it's a rhetorical question. Often there may just be salvage chemotherapy that is not gonna be successful. And so we could debate whether or not picking up a recurrence today is critical. It'll show up on the next scan. And that's some of that's philosophical. But it may not help the patient just in terms of survival, if we pick up the recurrence today, as opposed to picking it up on the next time, and to pick up post-treatment complications that may contribute to a decreased quality of life or actually threaten the patient's survival, even in the absence of cancer.

So when should we image, again? Eight to 10 weeks? What modality? I said that, what are the expected side effects that may be seen radiographically? Can we differentiate tumor recurrence from the post-treatment side effects? There there's occasionally overlaps and occasionally synchronous or coexistent problems. And there are endless pitfalls. So there's no way that I can get through every possible circumstance. It's a very complicated business.

Chemotherapy

But chemotherapy, lemme start with chemotherapy. Often given as an induction therapy that is prior to any other therapy for larger cancers, in order to shrink them prior to consideration for surgery, which is neoadjuvant therapy, when a patient cannot tolerate the toxicity of concomitant therapy, so that may be the recommendation, but the patient just can't handle the toxicity of concomitant therapy. They may get induction chemotherapy upfront and as salvage therapy after frontline therapy. It is also sometimes given as concomitant therapy that is simultaneous to XRT for larger cancers, cancers with more extensive nodal disease, patients, cancers for which surgery is not an option, not feasible, or the patient refuses. Okay. So you may have a patient who has a T four larynx cancer says, you're not taking my larynx out. That happens. Some patients just refuse to give up their larynx and then you have to do something else.

Concomitant therapy is superior to radiotherapy for advanced head and neck cancers in whom surgery is not going to be done. And post-op. For patients that are considered at high risk for distant metastasis, there is no perfect guideline radiographically, in order to determine that a tumor has responded to therapy, obviously you wanna see that the tumor is getting smaller, but there is no magic number. I would like to see at least a 50% shrinkage at six to eight or 10 weeks following completion of therapy. Obviously, if a tumor is enlarging, then your treatment is failing. When some abnormality remains, then there may be a role for PET in some cases to distinguish whether or not there is residual tumor or just the effects of therapy. It's not critical, as I mentioned, in every patient to find out that particular day, whether there's cancer, because if a patient's had induction therapy and the plan is to do radiotherapy, then obviously we'll image just to show that we're heading in the right direction. And we'll image again after radiotherapy. So it's not in every case in my practice to critically to know that day whether or not the therapy, the induction chemotherapy has resulted in a complete response, because we're gonna do other therapy and then mop it up.

So here's a patient with a T four maxillary sinus cancer. We can see an ugly, destructive cancer. It's busting out into the masticator space. And this patient was given chemotherapy. And the plan was they were gonna operate no matter what. But on this case, literally there was absolutely nothing as bad as that cancer was, there was nothing that I could point to and say, aha, I'm sure there's residual cancer, and no cancer was found in therapy. So this is an example of a cr or a complete response following chemotherapy.

This is a patient with a large and a deeply invasive oropharyngeal cancer who was given chemotherapy. And again, following therapy, almost nothing that could clearly be called tumor recurrence or residual tumor. One thing that I've noticed is that the worse a cancer is to begin with, the more likely that there'll be something bothering you on the scan after successful therapy. So even in the face of a curative therapy, the bigger and uglier a cancer was to begin with, the more likely there is gonna be some residual, some post-treatment residua that is gonna bother you.

Radiographically, here's a patient before and after three courses of induction therapy. You can see the medications here for supraglottic cancer. We see a very large and invasive supraglottic cancer. And then the bottom row matched images, virtually no evidence of tumor.

We are not at MD Anderson doing surgery for all larynx cancers. I know you didn't say that you should operate on every larynx cancer, but we actually operate on very few larynx cancers. At MD Anderson only really for salvage or for very advanced T four cancers.

Here's a patient with a left tongue-based cancer. The red arrows show you that cancer, this patient had induction chemotherapy, and the intent then was to have chemotherapy was to radiate thereafter. And there's often some small, do I know that there's cancer in here? So this is one of those situations where there might be a little bit of residual cancer. It may be that the cancer is cured, but there's a residual imaging abnormality. But this is one of those situations where it was not critical for me to know on that day, because the patient was going to get XRT subsequently. And I know that there's gonna be another round of imaging following that.

Here's a patient pre and post chemo acid. This patient had chemo rad for a right tongue-based carcinoma. And just to show you that really a very impressive response. I mean, the area of tumor has virtually no abnormal signal. But we often have to deal with areas of some persistent mucosal abnormality after radiation therapy. I'm getting ahead of myself just a little bit, but often there's mucosal irritation, mucositis, things that could confound, especially if it's exactly at the site of the original tumor. But this is the kind of scan that I would look at and say, I believe this to be a complete response. And we're not gonna image our patients one time after therapy and send them out into the world, right? Everyone's gonna get re-image. And so we learn to live with some uncertainty, unless it's something that we absolutely have to know.

Here's a patient with a right piriform sinus cancer. We can see the cancer here. This is before therapy, after induction chemotherapy. A nice response, after radiotherapy, there's some edema of the right supraglottic larynx. The AE folds a little swelled up, as we say, in the south. But there's really nothing here that would suggest tumor recurrence, as opposed to this patient. An example of where the therapy obviously failed, we had a T four laryngeal cancer. The patient refused surgery, he had induction therapy, and his bad cancer got even better. So this patient, I assume at some point, was compelled to have surgery.

Radiotherapy

Radiotherapy is the mainstay of head and neck cancer therapy. Most patients will, at some point in their course as head and neck cancer patients, get radiated. It's often the sole form of therapy for low volume oropharyngeal or larynx cancers, T one or T two exophytic cancers, T three tonsil cancers without extra capsular spread in the nodes. Many patients will also have XRT for high volume disease. After surgery, people who are deemed at high risk for local or recurrence will get radiotherapy.

The radiologists have to be familiar with the expected appearance of radiated tissue. And I'll show some examples as we go. There are the same issues of efficacy. How do we tell when a tumor is responding? It's getting smaller, and I'm gonna spend a lot of time at the end on the complications of radiotherapy.

I won't read this entire slide. Many of these things you are familiar with, these are the typical or expected radiographic findings after somebody who's been radiated, thickening of the platysma skin thickening, supraglottic laryngeal edema, retro pharyngeal edema, usually beginning within several weeks, peaking at several months, and variably. But over time, generally completely regressing.

How can I tell if therapy is working after radiation? No foolproof way. It's clear cut when it's obvious the thing melts away. But there's often middle of the road gray zones. And many of the things that I said with respect to chemotherapy apply. We may not necessarily have to know on that day. We may decide that we have to know, we need to get a PET scan. PET scan has its own pitfalls, pitfalls, I call 'em. But areas where you may be confused as to whether it's some effective therapy versus a residual or recurrent tumor.

But as I said earlier, if we pick up a cancer recurrence now versus picking it up on the next cycle, which may only be eight weeks away or 12 weeks away, we're not necessarily going to buy that patient a lot of time because the salvage therapies, once you've failed, especially if you've had chemoradiation and you recur, or the treatment fails, in most cases, the salvage is not very successful.

We can do ultrasound exams if we're talking about a nodal problem. And that's a whole talk in itself. The post-treatment lymph nodes, very complex subject. Sometimes a pet may be useful. We do a lot of ultrasound of the head and neck. And so in equivocal cases, we can always go back and do an ultrasound guided evaluation.

Here's a patient before and seven weeks after 57 gray for treatment of a pretty obvious tongue-based cancer here. And I would argue this as or submit this as a pretty good example of a complete response. There's nothing I can look at and say I think there's, I'm worried about it. You know, I can't even hallucinate something if I wanted to. This is just a stone cold complete response, at least on this day, pre and five weeks post 70 grade for a transo cancer. We can see the malignancy here. And afterwards, there's gonna be some fullness. There may be a little bit of enhancement, especially, it's very difficult. Radiated larynx can be very confounding. There can be areas of enhancement. It can be extremely scary, and you do your best. But in this case, I couldn't find anything that was convincing for tumor.

This is an example of somebody who actually reformed their thyroid cartilage after it was destroyed by a larynx cancer. So that's a good thing. If the laryngeal can cartilage just reform themselves out of what had been frankly destroyed bone or cartilage, that's pretty impressive. So that's something I'll look for and occasionally see as a sign that tumor that the treatment's been successful.

Here's a patient before, in about a month after induction chemotherapy. I dunno why they imaged that a month, but a very big cancer here in the tongue base given chemotherapy. Nice response to induction, chemotherapy and post-radiation. Notice that there is enhancement of the submandibular glands there. There's mucosal enhancement, a lot of typical changes that we would see after radiation therapy. But a good response, and nothing that specifically suggests tumor recurrence. And even the node has had a very impressive response, prior and after concomitant therapy for a massive really massive cancer. I suppose I should probably look up and see if this patient is still NED. This slide's probably a couple years old, so I wonder if this patient recured just between you and me, but a very, I think you would agree, a very impressive response to chemoradiation, nothing that is unequivocally tumor, bulky right-sided supraglottic carcinoma before and after radiotherapy. We see typical changes of radiation. There's a supraglottic edema. There's maybe a little bit of retro pharyngeal space edema here. But a very nice response to chemoradiation.

Sometimes we'll do PET scan. We don't routinely do PET scan in a lot of cases, 'cause I think the volume would just be too high. But this is the patient, this is actually the same patient who after this pretty nice response unfortunately did develop a paralytic space recurrence that was suspicious, certainly on CT and hot on pets. So just an example of failed therapy.

Here's patient before and after 72 gray. This is the failure slide. So we have a an anterior commissure glottic cancer here. And after, I don't know why five months went by here. So sometimes patients live somewhere else, and perhaps they don't come by. They don't return to us for follow-up imaging. And this patient's tumor obviously got big and necrotic. And this is just an easy example of treatment failure.

Complications of Therapy

So the rest of my talk has to do with complications of therapy, primarily radiation complications, mucositis isn't expected in a clinical finding. After radiation of therapy, they expect it, and it's a bad thing. It contributes to decreased quality of life for head and neck cancer patients. They're miserable. They can't swallow. They may need nutritional support, but it doesn't always have a dependable radiologic correlate. So just because I see mucosal enhancement doesn't necessarily mean that the patient has clinical mucositis, and they may have raging mucositis and not have very impressive findings. So that's not really a radiologic thing.

We can see enhancement in cystic transformation in preexisting thyroid gloss duct remnants, and various forms of necrosis are what I'll talk about. This patient was radiated for neuro pharynx cancer and actually did have a raging mucositis. And this is the kind of very striking mucosal enhancement that we may see after radiotherapy. But again, in this particular patient, he had mucositis, but it wouldn't necessarily correlate.

Several years ago, we published a series of patients that had, in retrospect, small thyroid gloss duct remnants, not necessarily full fledged cysts, but you look back and you could find small remnants. And it turns out that probably radiation therapy incites an inflammatory reaction in these remnants causing them to secrete and weep into themselves and get bigger. And here's three examples. This is somebody who, you never would've called this as a thyroid gloss duct cyst, I don't think. But it got a little bit bigger after radiotherapy. This one's a little bit more striking in its cystic transformation. And this one's still more striking. So this was a thyroid gloss duct remnant, you can see it going in a typical pattern of a thyroid gloss duct cyst. And this one underwent a fairly striking cystic transformation. So this is not a complication, perhaps, of therapy, but it's something that you might see and might find a little disconcerting.

Soft Tissue Necrosis

Soft tissue necrosis is a significant problem, even with conformal radiotherapy. We see a ton of treatment related necrosis. It is a huge problem in terms of quality of life for these patients. It can be if not understood as such, and I have numerous examples of this, where it's called tumor, it get biopsied and biopsy necrosis, you get more necrosis. Or if you get a PET scan, and the pet reader is a nuclear guy or somebody who's not tuned into this, it's hot because of the inflammation and the infection. It's a very easy misdiagnosis as tumor recurrence leading to biopsy, leading to more necrosis, and it can be very ugly.

So what we're looking for is air dissecting into the submucosal soft tissues in a place where air does not belong. And frankly, ulcerations. Now, some of these ulcers are easy to recognize as such, and some not. So I'll show you the spectrum.

Here's a patient who was in 2004. So he was radiated ending in March of 2004. So at this time, he looked pretty good. And then a few months later, along the lingual surface of his soft palate, there starts to be an ulceration. And then there's a bigger ulceration. This is very easy to miss. Okay? Now, I knew that in the mouth, by looking, the radiation oncologist was able to see an ulcer. This one's not the most obvious ulceration I've ever seen. This is more obvious after five months after radiation therapy for an oropharynx cancer. And here's an ulceration along the left pharyngeal wall. Imagine how painful that is, how difficult it is to eat. There's debris, there's air dissecting submucosal. There's probably a piece of the either the high, probably a piece of the OID that's getting ready to necros, and I'll come to that. But these require hyperbaric oxygen. And really it's hard to eat because your food gets trapped in there and it funks and it gets more inflammation, and it's just nightmare.

Here's a patient whose PET scan was read as concerning for local recurrence yet, except that on the ct, there's an ulceration. There's air submucosal in that tongue base. Okay? So this is not tumor. This is soft tissue necrosis.

Necrosis in Brain and Skull Base

We occasionally see necrosis in the brain in patients who are radiated for nasal pharynx or ear or parotid cancers, structures that are near the brain. This is a typical appearance of radiation necrosis in the brain. And somebody radiated for nasal pharynx cancer. It usually occurs in the anterior medial temporal lobe. It may be a little spotty areas of enhancement. It can progress. So very typical in patients with nasopharynx cancer.

We occasionally see skull base osteo radionecrosis and mandibular osteo radionecrosis is extremely common, especially in patients who have been treated for nasopharyngeal or oropharyngeal cancers in which the mandible is unavoidably affected by friendly fire, if you will.

Here's a patient who literally walked into our emergency room with altered mental status reporting that he had been radiated elsewhere for parotid cancer. And although that's the bone, windows are not the best, there's a disorganization of the bony architecture in the right temporal bone. And so basically, this patient had osteonecrosis of his temporal bone, which allowed air and organisms to get into his brain. And that's a brain abscess caused by osteo radionecrosis of the skull base.

A more typical example, this patient had been radiated for a skin cancer, and we have extensive disorganization and fragmentation of this patient's right temporal bone. And there's often some abnormal soft tissue enhancement adjacent to necrotic bone, which can be confounding and cause fear of cancer recurrence.

Mandibular Osteoradionecrosis

Mandibular osteo radionecrosis often begins as a lingual cortical defect, the kind of thing you see here. And over time, it can progress to pathologic fractures, lytic changes, and frank breakdown of bone. And sometimes it can lead to infection in the masticator space. So this is a patient who ended up with a masticator space abscess secondary to his mandibular osteo radionecrosis. Just another case which led ultimately to a pathologic fracture in the mandible. So this is advanced mandibular osteo radionecrosis.

This patient was radiated for a base of tongue cancer, and the PET scan was hot in this location. And the pet reader, a nuclear person, didn't call it cancer, he said he thought it might be inflammation or something, but he also said that there was no CT abnormality on the accompanying CT images. But there actually is, 'cause this is what the mandible looked like in November of 11. And now, although it's on the buccal side, not the lingual side, there is a cortical defect in that bone. And this is a scan I read just the other day, and now there's a huge sequester of dead bone in that mandible. And the reason it was hot on PET is because this is an act of inflammatory process. This is not just dead bone sitting there in an inert fashion. This is inflamed, infected bone, and it's usually gonna be very hot on PET scan. And that's what we're seeing here. But you wouldn't wanna see that called tumor.

Hyoid Osteoradionecrosis

Osteo radionecrosis of the hyoid bone. I'm guessing that it's underappreciated. We published a series of patients in the age a NR. It can be confused with or coexist with recurrent tumor. And what we see in our cases is a defect in the oropharyngeal mucosa down in the molecular in which air dissects in sub mucosally and comes into contact with the hyoid bone. And we're also looking for abnormal architecture of the bone. We're looking for air and fragmentation within the hyoid bone.

Couple examples. Here's a patient that began after chemotherapy and radiation with a small air-filled defect in the ula. This continues to grow, and ultimately we have air surrounding the hyoid bone. You're not supposed to have air around your bone, you're certainly not supposed to have air in your bone. Air and bone is a bad thing. That's my rule of thumb. Air and bone is bad. You can take that to the bank. That's never normal here so that the hyoid is crumbling. And here's a patient who has a really advanced case of hyoid osteo radionecrosis, and unfortunately, massive tumor recurrence as well. So you can occasionally have coexistence of these treatment complications.

Here's a patient with a tongue-based cancer. We can see it. And at the time of the treatment, the hyoid bone looked fine. And then after a while, now there's air dissecting into the soft tissues. There's areas of metabolic abnormality at the site. And that is a raging case of od osteo radionecrosis, no tumor. But you could see that the PET scan is mighty scary, and could easily be called tumor.

Laryngeal Osteoradionecrosis

Laryngeal osteo radionecrosis is a significant problem in patients who are treated with non-surgical therapy for their cancers. Although it can occur, it can coexist with recurrent tumor. It may also require, and he, a Dr. Aria referred to this, A patient may actually have a successful therapy for their larynx cancer, but have a deconditioned and non-functional larynx that nevertheless requires laryngectomy because of things like Conrad necrosis.

So air dissecting into the soft tissues where air doesn't belong, bone surfaces or cartilaginous surfaces exposed to air and fragmentation with pathologic fracture. A couple quick cases, and I think I'm just about done. This patient was radiated for a hypo pharyngeal cancer because of, quote, an abnormality on PET ct. The patient had three biopsies. So you biopsy necrosis, that's bad. You get more necrosis. We can actually see air dissecting into where the retinoid should be. And this patient actually has air in the retinoid air and bone is bad. And now the patient is aspirating. We can actually see the aspiration on this esophagus. There's some debris and crapola, if you will, on the surface there. And over several months time, we can see that this patient has essentially auto amputated the hyoid bone. So I've seen this on a number of occasions in which we had a hyoid bone, and then it didn't look so good. And now there's air and debris where it used to be. And there's no more retinoid. I assume he swallowed it or aspirated it. I hope he swallowed it.

So this is another patient who was radiated elsewhere for a supraglottic cancer, now has strider. And we see air dissecting essentially through a big defect in the thyroid cartilage. Now you would be hard pressed to exclude cancer recurrence here, except to say that there's really no mass, there's no abnormal enhancement. So if you see no abnormal enhancement and a defect in the cartilage is a reasonable bet that what you're looking at is chondral radionecrosis, and not tumor recurrence.

So here's a patient with the superior corn U of the thyroid. Cartilage is exposed to air. There's air bubbles. It's starting to fragment. Another typical case of chondral necrosis. Another bad case, this might be my last case for this meeting, three in 10 months, post the XRT for a base of tongue cancer. This is some of the findings you might see on a PET scan that would make evaluation post therapy more difficult. The patient had some mucositis. We can see enhancement, and now it looks like on the PET scan, because it's there's a shine through. So the abnormality on a PET scan will often look bigger than it is in real life. And consequently, this whole thing looks like a big mass on PET scan, but really there's no mass there. That's just mucositis. And often we'll see areas of metabolic activity like in the tongue, especially in the tongue, that don't really have a corresponding abnormality. And that's not cancer, that's just a patient moving their tongue around. So there's a lot of things on PET scan that can be confounding after therapy. And this tongue was in fact soft. But I'll tell you that, that's a scary looking finding on PET scan.

Conclusion

To conclude, post-treatment imaging is tough. There's no substitute for experience and unfortunately for errors. So what I try to do is maintain a close contact with my clinicians, get as much information about the patient's treatment and their condition as I can, and try to just try to get better and show my mistakes to others and get better together. That's what I do. And thank you again for involving me in this meeting.