A website dedicated to better

Zach Berman MD

Y9004.PreTreatmentPlanning

Lung Shunt Calculation

To conclude my past few weeks of re-thinking the lung shunt fraction, I was hoping to take this week and maybe be a little more practical – because we do I have to offer something in return.

So, if MAA is not a perfect (or even plausibly useful) marker, then what can we do. I would argue that we should use the actual radioembolic product itself. Now, this isn’t a super novel idea. In fact, there is even a product on the market that is good for both imaging AND treating. Ho-166.

Which has not only the β decay, but it also has a γ decay, making it ideal for SPECT imaging. This imaging can be performed not only during the mapping process, but can also be used during the treatment process.

A great example of this, was the 37 patient cohort published by Wagemans et al. (1) They found that the estimated lung dose was going to be 1.53 Gy  (0.09–21.33 Gy) with MAA vs 0.00 Gy (0.00–1.20 Gy; p < 0.01). Now, one might say, its only a 1 Gy, difference, how bad could that be. Well, take a look at the extremes. The MAX dose absolute difference was 20 Gy!

More importantly, they found when looking at the actual treatment images, that Ho mapping and Ho-treatment were significantly correlated while there was no correlation between MAA and actual treatment lung doses.

(example from citation #1 showing the very large differences between scout MAA, scout Ho, and Ho treatment)

Well, one might ask, if I like the US, then I don’t have access to holmium (yet) is there anything I can do? The answer is, maybe, someday, as shown by Kokabi et al. (2), where they used a scout dose of Y90 RESIN microspheres in 30 prospective patients and compared it also to MAA. They found only a weak correlation between MAA and SirSphere infusion, and a strong correlation between scout doses and Y90 infusion.

(example from citation #2, showing the very large variation with MAA not really showing a relationship and the very high correlation with scout Y90.

So where does this leave us. We examined what it meant to be a “high” lung shunt and where our dose limits came from. We looked at the issues with calculating LSF with planar imaging only, and now we have posited the benefits of using the actual radioembolic device for mapping and lung dose calculation. All is good now, but (at least in 2024), how do I think we should use LSF?

Practical thoughts:

  1. Push Y90/Ho166 companies to do the work to make “scout” doses the standard of care. Not only on the research side, but also on a commercial side. (e.g. if I wanted to do a scout dose today, there is no feasible way for me to do it given commercial offerings of their products)
  2.  If using MAA, use SPECT imaging over planar imaging. Follow the 30/50 rule but remember where it came from and the rarity of radiation pneumonitis. We need better metrics, but it may take a multi-institutional VERY large cohort to figure out.
  3. To continue our pondering of what lung shunt fraction (LSF) is, and how we consider it in our decisions on dosimetry for radioembolization of hepatic tumors, I want to turn our attention to how LSF is calculated.
  4. When we look back at our previous thoughts, we see that initially the single anterior planar view was used to obtain the 30 Gy and 50 Gy, thresholds we use today. What does that look like? See Figure 1. A very crude image is drawn to estimate where the lungs are, as well as what the liver looks like. Some companies will even do the contouring for you with the benefit of “AI”.
  6. Fig 1: A typical planar LSF calculation
  7. Then a simple percentage is taken, multiplied by 100, and voila you have your number. (Eq 1)
  9. Eq 1. LSF simple calculation
  10. Well, people started to think, could we do this more intelligently. What if we got both anterior and posterior views and took the geometric average. (Eq 2)
  12. Eq 2. The LSF geometric mean calculation.
  13. More complex, but ok. What’s the difference? Looking at 30 patients Erwin et al (1), found that the differences were “significant”. And by that, I mean statistically, in reality the absolute difference was about 1%. Such a difference would change dosimetry for your procedure in only a minimally meaningful way – especially not to the point of making us have to take a square root! (which I suspect none of us have done since high school)
  14. Though, as technology has increased, many have turned to using SPECT/CT, mainly as a source to be able to calculate a tumor-to-normal ratio for partition dosimetry. But that, will have to be a discussion for another day.
  15. The SPECT can also be quite helpful for LSF, as it has better resolution and ability to quantify (with some corrections) a “true” LSF. When comparing SPECT/CT to Planar, there have been a few people who have looked at this. The first to look at would be Dittman et al (2). They looked at 50 patients and found that the median LSF was 6.8% on planar imaging and 1.9% on SPECT/CT. Now, that is more clinically significant difference. They go on to note:
  16. “On the basis of planar imaging, dose reduction or even contraindications to SIRT had to be
    considered in 10 of 50 patients…in contrast, SPECT/CT quantitation showed substantial shunting in only 2 of the 50 patients.”
  17. Interestingly, they observed that 13% of the total MAA signal came from places outside of the liver or lungs. In this case, most likely free Technetium– which plagues me. I can see thyroid, stomach, and kidneys on most of my cases, which makes me wonder how accurate MAA is compared to the radioembolic itself. But, alas, this too, will have to be for another day.
  18. But, you would say, those were only 50 patients. Elsayed et al. (3), found quite similar numbers in 293 consecutive patients. Their mean planar LSF was 8.27% compared to SPECT LSF, 3.27% (p < 0.001). Interestingly, the same absolute change of 5% was noted in both of these studies. Hard to know what to make of that, but it’s interesting, nonetheless.
  19. They also found that patients with planar LSF ≥ 20% had a greater absolute discrepancy 13.31% vs 4.74% for those less than 20% (p < 0.0001). For a change that large, you might even cancel a procedure based on the planar LSF altogether.
  20. Young et al (4), specifically looked at these patients who had higher lung shunts. In their study, of shunts greater than 7.5%, they found that mean planar LSF 14.7 and mean SEPCT LSF 8.7 were quite different (P < 0.001). In this case 6%, quite like the previous two reports of 5%…the trend continues. Looking at their cases, 24.1% would have required dose reductions or cancelation if planar imaging was to be used alone. No bueno.
  21. Though, those are just percents, what does this even mean for DOSE. As again the 30/50 Gy paradigm is a dose not a percent. Well look at Eq 3 to see how we convert.
  23. Eq 3: LSF to lung dose calculation
  24. Lung mass is always assumed to be 1 Kg. But clearly, not every person’s lungs are the same. Interestingly the 1 kg mass is from the International Commission on Radiation Protection (ICRP) Publication 2348. ICRP Publication 8949 updated that value to 1.2 kg for men and 0.95 kg for women. As a mean, that’s already a 20% difference for men and a 5% difference for women. But not everyone is average and could have substantially more or less depending on the actual lung mass of your patient. This can wildly change what our estimated lung dose actually is. (5)
  25. This is not even considering that the lung dose is evenly distributed. As our pulmonology colleagues will remind us, most of the blood goes to the lower lobes than the apices. Therefore, an assumption that lung dose is equally distributed across the assumed 1 Kg, can’t be true. Some areas would be hotter with more dose, and others colder.
  26. I think it’s clear, if we are to believe the 30 Gy/procedure 50Gy/lifetime doses. We NEED to do better with both % calculation which can be significantly different based on planar vs SPECT or something as simple as how reliable your MAA even is.  Even with perfect MAA – does that predict the true radioembolization lung dose? Something to explore next time…as people smarter than me, have already looked into this.
  27. ____________________________________________________________________________________________________
  28. (1) “Anterior-only versus geometric mean 90Y microspheres lung shunt estimation: Implications for in-room, single-view gamma camera imaging” Erwin W, Mar M, Kappandath S. Journal of Nuclear Medicine May 2013, 54 (supplement 2) 2079
    (2) Dittmann H, Kopp D, Kupferschlaeger J, Feil D, Groezinger G, Syha R, Weissinger M, la Fougère C. A Prospective Study of Quantitative SPECT/CT for Evaluation of Lung Shunt Fraction Before SIRT of Liver Tumors. J Nucl Med. 2018 Sep;59(9):1366-1372. doi: 10.2967/jnumed.117.205203. Epub 2018 Jan 25. PMID: 29371406.
    (3)Elsayed M, Cheng B, Xing M, Sethi I, Brandon D, Schuster DM, Bercu Z, Galt J, Barron B, Kokabi N. Comparison of Tc-99m MAA Planar Versus SPECT/CT Imaging for Lung Shunt Fraction Evaluation Prior to Y-90 Radioembolization: Are We Overestimating Lung Shunt Fraction? Cardiovasc Intervent Radiol. 2021 Feb;44(2):254-260. doi: 10.1007/s00270-020-02638-8. Epub 2020 Oct 1. PMID: 33000319.
    (4)Young S, Flanagan S, D’Souza D, Todatry S, Ragulojan R, Sanghvi T, Golzarian J. Lung shunt fraction calculations before Y-90 transarterial radioembolization: Comparison of accuracy and clinical significance of planar scintigraphy and SPECT/CT. Diagn Interv Imaging. 2023 Apr;104(4):185-191. doi: 10.1016/j.diii.2022.12.002. Epub 2023 Jan 3. PMID: 36604211.
    (5)Busse NC, Al-Ghazi MSAL, Abi-Jaoudeh N, Alvarez D, Ayan AS, Chen E, Chuong MD, Dezarn WA, Enger SA, Graves SA, Hobbs RF, Jafari ME, Kim SP, Maughan NM, Polemi AM, Stickel JR. AAPM Medical Physics Practice Guideline 14.a: Yttrium-90 microsphere radioembolization. J Appl Clin Med Phys. 2024 Feb;25(2):e14157. doi: 10.1002/acm2.14157. Epub 2023 Oct 11. PMID: 37820316; PMCID: PMC10860558.
  29. The lung shunt fraction (LSF) estimation has long since plagued a lot of interventional oncologists seeking to treat liver malignancies. Often, the digital subtraction angiogram (DSA) looks great, as does the cone beam CT (CBCT), but unfortunately, the lung shunt estimation – usually from planar scintographic images – is too high to tolerate adequate treatment due to excessively high lung doses (30 Gy single treatment, 50 Gy life time treatment per IFU on both glass and resin microspheres) which can lead to radiation pneumonitis (RP). At least this has been my issue in many cases; below is an example of case that I had with no obvious macrovascular shunting but an MAA scan with an 30% shunt. The choice is to still treat with significant dose reduction or consider other means?
  31. So this got me thinking, a) what is the history of LSF and b) what is the current evidence for it? I had a vague understanding that it was external beam radiation data that biased the limits, and to some degree, I was correct. Though, this rabbit whole goes a lot deeper, and it is probably about time we start to think more critically about the role of LSF to adequately dose tumors to get optimal response, as well as continue to stay within safe realms as RP is a real event, albeit quite rare.
  32. To start off this adventure, I came across this great review article, that I recommend most read if you are interested in topic as they break it down a lot more eloquently than I can here. But to summarize, all of our dose limits come from TWO PATIENTS- yes, only 2 patients.
  33. The two papers that first looked into this were Leung et al. (1995) Ho et al. (1997) (1,2). In the first case, where the dose limit of 30 Gy / treatment likely came from showed that radiation pneumonitis developed in 1 of 3 total patients who had a lung dose of over 30 Gy (32 in their case) and in only 1 of 95 patients in general. To make things even worse, they only calculated the dose on the single anterior planar image – though planar scintigraphy and SPECT calculations will have to be left for another day.
  34. The second paper, Ho et al. showed that in one of two (1/2) patients that had a life time dose greater than 50 Gy develop RP. These patients received 54 and 59 Gy. Hence, the likely origin of the 50 Gy lifetime dose limit.
  36. From Leung et al, diagnosis of RP.
  37. An interesting historical note. As I was taught, these numbers came from “EBRT data” from lung irradiation. This is possibly true. Dating back to at the 1960s, whole lung radiation was used to treat metastatic tumors. In a brief series published in Radiology in 1968, authors suggested that a 2500 Rad limit, seemed reasonable. This would be 25 Gy with the more commonly used unit today. Though, this threshold also was developed on a limited number of patients possibly plaguing radiation oncologists to come similar to the radioembolization thresholds of today for radiologists. (3) Parenthetically, I suppose at the time, radiologists and radiation oncologists were one and the same as they had not split by then.
  39. How have we evolved? Two new studies have looked further into this Salem et al (2008) (4) and Das et al (2020) (5). In the former, 58 patients were evaluated with lifetime doses greater than 50 Gy. In this cohort, the 18 patients that received a single treatment lung dose >30 Gy (average 37.1 Gy), none had RP. None of the 58 patients with high lifetime doses had RP.
  40. Das et al. [same group/institution] expanded upon this with a larger cohort (103) of patients with a LSF > 15% and showed that outcomes were unchanged, and no clinically significant RP events occurred. While the median, 22 Gy, lung dose / session was below the 30 Gy recommendation, maximum single session dose was up to 52 Gy in the cohort.
  41. Lastly, for more thought provocation, Stella et al (2022) (6) showed that in their total cohort of over 300 patients, 18 had an estimated lung dose of greater than 30 Gy based on the mapping MAA. Though, on post Y-90 PET, none did. Two patients had RP. One with an estimated dose 89 Gy from the mapping and the other with a dose of 34.1 Gy. They note, no radiation pneumonitis developed in patients with a measured PET dose lower than 12 Gy.
  42. Though my caution would be to not fall for the same trap again – as the number of patients with RP was too low to have practice changing effects. Future work is needed, given the rarity of RP, likely on a multi-institutional bases. I can dream at least of this one day happening…although I won’t be waiting with baited breath.
  43. My take from this adventure, lung dose thresholds of 30 Gy single procedure, 50 Gy lifetime are relatively arbitrary based on essentially two patients in two case series over 25 years ago and/or EBRT data dating back 56 years go. Radiation pneumonitis is real, but likely less common than anticipated.
  44. More work NEEDS to be done, since as Das et al showed, dose reductions are happening due to these thresholds that may under-dose the tumor to levels not adequate get response.
  45. That being said, I am open to changing my mind on this subject.
  46. ___________________________________________________________________________________________________________
  47. (1) Leung TW, Lau WY, Ho SK, Ward SC, Chow JH, Chan MS, et al.. Radiation pneumonitis after selective internal radiation treatment with intraarterial 90yttrium-microspheres for inoperable hepatic tumors. Int J Radiat Oncol Biol Phys. 1995; 33:919–924
  48. (2) Ho S, Lau WY, Leung TW, Chan M, Johnson PJ, Li AK. Clinical evaluation of the partition model for estimating radiation doses from yttrium-90 microspheres in the treatment of hepatic cancer. Eur J Nucl Med. 1997; 24:293–298.
  49. (3) Margolis, Lawrence W., and Theodore L. Phillips. “Whole-lung irradiation for metastatic tumor.” Radiology 93.5 (1969): 1173-1179.
  50. (4) Salem R, Parikh P, Atassi B, Lewandowski RJ, Ryu RK, Sato KT, et al.. Incidence of radiation pneumonitis after hepatic intra-arterial radiotherapy with yttrium-90 microspheres assuming uniform lung distribution. Am J Clin Oncol. 2008; 31:431–438. 
  51. (5) Das A, Riaz A, Gabr A, Ali R, Mora R, Al Asadi A, Mouli S, Lewandowski RJ, Salem R. Safety and efficacy of radioembolization with glass microspheres in hepatocellular carcinoma patients with elevated lung shunt fraction: analysis of a 103-patient cohort. Eur J Nucl Med Mol Imaging. 2020 Apr;47(4):807-815.
  52. (6) Stella M, van Rooij R, Lam MGEH, de Jong HWAM, Braat AJAT. Lung Dose Measured on Postradioembolization 90Y PET/CT and Incidence of Radiation Pneumonitis. J Nucl Med. 2022 Jul;63(7):1075-1080.