Cause Of Liver Cancer

thank you. so, this is day 2 of our courseand i have been given the challenge of talking about imagining image guidance in sbrt. ifany of you have started to treat liver cancers, you have realized that imaging is a crucialpart of the process for any radiotherapy treatment. hopefully at the end of the seminar, i cantell you that you don't need the best imagination if we have the tools that enable us to actuallysee the tumour. i'll first acknowledge many people who have been a part of the liver radiotherapyprogram at pmh in the past 10 years and thank you in advance for all your contributions.i am going to outline why focus on liver cancer, why sbrt, what's the rationale for igrt, andbriefly summarize some igrt approaches and some selected outcomes. hepatocellular carcinoma(hcc) is an important cause of global cancer

death. it is the 5th cause of cancer, 3rdcause of cancer death and unfortunately, 5 year survival rates are still very poor(lessthan 10% internationally). it is the number one cause of increasing cancer death in theusa, so becoming much more of an important problem in north america. many patients withliver confined disease progress in the liver with a vasculature beside the liver and it'sa local regional problem. many patients are unfortunately unsuitable for standard localtherapies and radiation has a potential large role to play in the treatment for many patients.metastatic colorectal cancer is also an important disease. the survival of patients with metastaticcolorectal cancer is improving; median survival has doubled over the past decade. reasonsfor improved outcomes include improved systemic

therapy but also increased use of surgeryor local therapies for liver metastases. colorectal cancer metastases are cured and in some patientswith resection, 40 to 50% 5 year survival in most series. but many patients again endup with liver confined disease either refractory to surgery or other local therapies or patientsare unsuitable for surgery due to medical contraindication or surgical challenges. so,what about non colorectal patients? there is definitely a boom in interest of treatmentof oligo metastases. this was described in the 1990s as an intermediate state of cancerbetween localized and widespread metastases representing a low burden of disease, potentiallycurable with local therapy. in this very nice review, jco selected outcomes with surgeryrfa or sbrt are described for not only liver

metastases but other oligo metastases in adrenalglands, nodes and in the lung. cures have been seen following local therapies for oligoliver metastases from not only colorectal cancers but breast cancer, melanoma, renalcell, sarcoma, lung cancer, so it's a really interesting arena of study.so, why sbrt? first state, there are clearly many ways to deliver a radical dose of radiotherapythat may ablate cancers. external beam radiotherapy techniques are perhaps the most widely available.brachytherapy requires specialized expertise. there is very interesting data coming fromgermany but that is clearly not something that is widespread. intraoperative radiotherapyprobably has only very few indications maybe in the operative settings. the radioisotopesinteresting again requires expertise and it

is an invasive approach. so, of the externalbeam techniques again, ions, protons not available, high cost. the radiotherapy approaches hypofractionationis perhaps something more feasible for most of us. so, what is stereotactic body radiotherapyor stereotactic ablative radiotherapy, sometimes referred to as sabr. very conformal dose distribution,highly potent dose, high dose per fraction, motion management requires image guidancea stereotactic component of the sbrt and it generally refers to few fractions of radiation3 to 6, in the us the billing code is 5 or fewer. with fewer fractions, it is more convenientfor patients and improves throughput. but these same techniques can in fact be usedin whatever fraction is best radiobiologically for patients. is there a biological rationalfor sbrt? well there are high dose per fraction

effects. when the does per fraction is 8gyor higher there are some effects that have been seen that are not seen at standard fractionationand some of these effects are summarized here. injury to the endothelial cells that are supplyingthe tumour.and two mechanisms of injury that again appear more common more likely witha higher dose per fraction. immune phenomenon are seen again with more likely a higher doseper fraction and all of these could lead not only to local control but potentially theabscopol effect where the distant disease outside the radiation field could in facthave a response even with a local therapy so very exciting area of research.there are some downsides of sbrt. sbrt is more sensitive to uncertainties than conventionalradiotherapy that is geometric, dosimetric,

contouring uncertainties. the planned radiationtherapy doses are not the delivered radiation therapy doses. this is of course any treatmentand any radiation plan but as we increase the dose per fraction, increase the conformality,increase the dose gradient, we are more sensitive to these uncertainties. systematic errorsin positioning could lead to a geographic miss, excessive toxicity and/or reduced tumourcontrol. so, why igrtlet's first start by borrowing a slide of a diagnostic multiphase ct for a hepatocellularcarcinoma patient; this is from dr. jang a radiology colleague. if you look at the hepaticarterial phase in the left and the portal venous phase in the right, you can see thereis tumour in the portal vein. but of note,

you need to see the delayed phase of imagingto actually appreciate this is a very very extensive tumour. so, you could imagine ifwe did not have appropriate imaging and delineated a small tumour, most of this tumour wouldbe missed with radiotherapy. so for hepatocelluar carcinoma, multiphasic imaging is crucialfor contouring. as one of my colleagues likes to say "garbage in, garbage out." so we couldhave state of the art image guided tools but if we don't have appropriate imaging and contouringthen those tools won't have t the effects we want.this is an example of a typical sbrt plan. this is a hepatocellular carcinoma with invasioninto the portal vein. being treated with 35 gy in 5 fractions is shown with the pink isodose.most often in our clinic, tumours like this

are limited in the dose that we can deliverby normal tissue tolerances. either the volume of liver irradiated or perhaps even more commonlyproximity to luminal gi structures such as the stomach, duodenum, and small bowel. sothere is a strong motivation for reducing the ptv and we need advanced technology andigrt to do this. in this particular example you can see it is the luminal gi structuresthat limit the dose for this particular patient and with a reduced ptv we may in fact be ableto give higher doses to the actual tumour. the clinical rationale for igrt, dose responseleads to improve tumour control. but unfortunately, higher doses also lead to an increase riskof toxicity. a igrt can improve the concordance between planned and delivered doses. so, itwould increase the chance of uncomplicated

sustained local control.unfortunately, regardless of the motion management strategies that we use to deal with breathingmotion, shifts in the liver position relative to the vertebral bodies occur. this occursin patients treated with breath hold, free breathing, gating, compression and tracking.so imaging at the time of radiation delivery, igrt is needed to account for these shifts.this is some data from two experiences looking at breath hold reproducibility in michiganand toronto. the standard deviation of that distribution of shift of liver relative tothe vertebral bodies. you can see that there is a larger variability in day to day repeatbreath hold then within the time period of a treatment fraction meaning that daily igrtis required not just one day at the beginning

or just selected fractions. this is data frompatients treated with free breathing or abdominal compression. you can see the intrafractionchange in the liver position relative to the vertebral body on the y --axis and you cansee the intra. sorry, so inter-fraction to fraction variability on the left and intra-withinthe time of a treatment fraction on the right. so you can see between fractions occasionaloffsets of a cm are seen for some patients. whereas, in the time period of fraction, inour own experience for treatments 30 minutes are less, the majority of change is 3mms orless. but you know one outlier, in this patient in fact had capsular metastases causing painand interestingly after his first two treatments when pain responded, he had improved reproducibilityand less intrafraction variability for the

remaining treatment.so, what are some approaches we can use to image guide and treat these tumours appropriately?there is a plethora of technologies, all have pros and cons. this is summary of some ofthe tools available. there are mr image guidance systems in development and combined systemswhich are now in use. just a summary that liver igrt is challengingas mentioned, we need iv contrast to see the tumour. motion due to breathing and stomachfilling has to be considered, shifts in liver position as i mentioned and organs can deform.there are many tools so it is hard to know what tool may be best for certain situations.of importance we have learned a lot with our image guidance tools about tumour and organmotion, deformation and unexpected changes.

we need to have a surrogate for the livertumour and at the treatment console that may be any of what is listed here. that may bethe whole liver, portion of the liver, portion of the liver, lipiodol following tace, fiducialmarkers, calcification, clips, maybe even iv contrast at the time of imaging at thetreatment unit. intfrafraction solutions so solutions forcontinuous or frequent imaging during one fraction are most important if there is longanticipated treatment times or we anticipate shifts due to pain or movement disorders.solutions are listed here: frequent or continuous imaging with fluoroscopy for example, opticalmonitoring, transponders or other solutions that can be complimentary to 3d imaging.this is something that isn't wide stream and

we in fact do not use this routinely now butbefore we went to an igrt approach, we did use this routinely at our units. sometimesthe beam itself may allow some soft tissue surrogate for the tumour that may be visible.in this patient for example, you can see the air-diaphragm interface to have some typeof reassurance as the beam is actually being delivered that you are in the right spot.real time tumour tracking is an option where the actual gantry will move with the patientas they breath. tumour gating is when the beam is on during only one phase of the breathingcycle for example exhale. this is an approach initiated in japan in 2000 by the kyoto group,now commercially available with unit gimbal that can move and in fact also do cone beamct.

interfraction igrt solutionsas i mentioned before the shifts or the day to day setup changes are the largest sourceof potential systemic uncertainty. so, correcting those shifts. in liver position most oftenmost marked in the superior inferior cranial-caudal direction is very important. planar/fluoroscopicimaging can be complementary to 3d volumetric imaging.this is an example of a poor's man igrt, mv orthogonal imaging. this is what we startedwith in 2003 in toronto before we had volumetric imaging or orthogonal kv imaging at the unit.the diaphragm was used as a surrogate for the liver tumour and vertebral bodies to alignpatients. simple using older technologies, the drr and mv portal images matched. it cancorrect for some of the offsets in the tumour

position. as we transitioned to cone beamct, in our first group of patients there was guidance using the mv orthogonal diaphragmapproach but we obtained cone beam ct in these patients. so it allowed us to then look packand say how was the liver actually positioned using that type of old man igrt. you can seebefore any type of repositioning, there are offsets of 20mm in some patients' in the liverposition itself. but using the mv guidance approach, the great outlier were correctedfor and in fact the majority of the offsets were within 8mm but one third of offsets intotal liver position were more than 5mm in size. so, clearly for small targets, highdoses in lesions in a deformable liver we want to do better than that. so, orthogonalkv fluoroscopy is one approach. this can allow

confirmation of liver motion and breath holdbefore every fraction and it can correct for large crania lcaudal offsets as i described.but even better is inserting a fiducial marker around the tumour. we will hear more laterabout some strategies to do that. 3d volumetric imaging allows us to see notjust necessarily the liver but also adjacent organs such as the stomach which may be doselimiting and lipiodol following tace is very useful for hepatocellular carcinoma igrt.volumetric allows quantification and planning for rotation and deformation of the liverand adjacent structures. it allows for visualization of unexpected findings and determination ofaccumulated delivered dose vs. planned. i think this is an important research tool wherewe really know what doses we are giving then

we are going to have a better idea of whatthe true dose volume relationship is for normal tissue complications and tumour control probability.this is an example early on of an offset that we saw that wasn't necessarily anticipated.due to the baseline shift in liver position relative to vertebral body, when the patientwas treated with cone beam ct and there is liver image guidance. these contours are fromthe planning ct and they are overlying the cone beam ct. so the liver matches quite nicely.but if you can actually see, the spinal canal is actually much closer to the liver withthat baseline shift compared to the blue counter that shows its position at planning in factwithout accounting for that could of lead to a higher dose than anticipated to the spinalcord. so now we in fact use use prvs... step

dose gradient away from structure that weanticipate will move during treatment. another example of how volumetric imaging showed usa detection of a change in liver volume. this is in a clinical trial where patients weretreated with 7 days of sorafenib (a tyrosine kinase inhibitor). we found unexpected livershrinkage and this was something that hasn't been reported, we did not expect it. we wouldnot have known about it if we had used a fiducial type of image guided approach.so what about breathing motion? how do we consider that? this is a patient treated in2005. you can see his cone beam ct during free breathing has artifact a little bit difficultto guide off. this same data set when only those projections are taken to reconstructthe cone beam from exhale, you start to see

a little more clear that the liver may beuseful but clearly if that same patient is imaged and treated in breath hold as you cansee it is much easier to anticipate how this liver can be used for guidance in the freebreathing. we have solutions for all these situations but the ones with 3d cone beamct in motion are challenging. respiratory sorted cone beam ct is now available commercially.it allows automated breathing amplitude measurement and matching. and as i mentioned, i thinkof the most simper solutions, stop and go cone beam ct allows simpler manual matchingor automated matching and easiest verification it is something we use routinely used in ourclinic. this is an example of a non-sorted cone beamct and with respiratory sorting you can also

visualize the breathing and motion of theliver that day immediately before treatment seen here. the yellow is the ct at planningat exhale, the red is in inhale. so if you look closely you can see that at the treatmentunit that the breathing is similar to how it was at the time of planning.so we're going to change gears and just briefly focus on some selected outcomes. this is alist of some of the potential radiotherapy toxicities that are crucial to avoid. radiationinduced liver disease can be life fatal. classic radiation induced liver disease consist ofanicteric hepatomegaly, ascites and elevated liver enzymes and that's more common in metastasesso we have a much better understanding now about how to avoid this. but non classic radiationinduced liver disease is less predictable.

and more challenging to avoid that may includeanything form elevation of transaminases, reactivation of hepatitis, a general declinein liver function, thrombocytopenia. also, we need to consider the common bile duct andluminal gi structures and chest wall pain and rib fractures. there is a growing bodyof literature to give us advice about dose volume constraints and limits for all of thesestructures. this is an example of some dose volume constraintsrecommendations for liver to avoid liver toxicity. liver toxicity is more common in patientswho have underlying liver cirrhosis or have viral hepatitis b. as the underlying liverfunction is worst the risk increases. liver metastases guidelines as proposed by the coloradogroup, if you can save 700 cc to less than

15 gy in 3 fractions is something used routinely.for hepatocellular carcinoma there are different strategies to avoid toxicity. i'll highlightone recommendation, looking at maximizing the volume of liver that is spared for radiotherapy.even in patients where 700cc receives less than 18gy there is a 11% risk of decline inchild-pugh score and very steep increase in risk is that spared liver volume decreasesmodestly. so, really we need to maximize or decrease the volume of liver that is treatedwith radiotherapy. so, the splash of low dose is probably important.selected outcome: rfa vs. radiosurgery. this is retrospective but intriguing that thereis significantly longer disease free survival compared to rfa 34 months vs. 6 months forcolorectal cancer and there is a dose response.

this is a pooled analysis suggesting a doseof 48-51 gy in 3 fractions for 1 year local control of 90%.looking at accumulated dose with the volumetric image guided information that we have. asi mentioned we might have a better idea of the true dose to the tumour. this is workdone by mike velek, swaminath and kristy brock. you can see when we look at the accumulateddose; there is a bigger spread in terms of the hazard ratio for dose response. so, veryinteresting. hepatocelluar carcinoma is a responsive tumourand this just shows a quite large tumour treated with modest dose of radiation (33gy in f fractions)with a sustained response. probably not ablated but by controlling this tumour for some timewe can hopefully impact and improve survival

for that patient.a more recent series from japan for hepatocellular carcinoma over two hundred patients mostlysmall tumours at 84% t 1., mostly child a, again using relatively small doses. 35 to40 gy in small fractions. three year local control and survival of 91 and 7 %, outcomessimilar to rfa and surgery. there is a suggestion of dose response andthis one recent paper that reports outcomes in slightly more advanced liver cancer patientsusing 3 fractions 33 to 60gy. you can see the improved survival and local control whenhigher doses were delivered. this was independent on a multivariate analysis; however therewas a poor relation between the volume of tumour and dose. so, the higher volume lesionstended to get lower dose.

radiation can also effectively treat hepatocellularcarcinoma with portal vein thrombus. this is an example of a patient with relativelymarked disease and here is an example of the tumour invading the portal vein, very highalphafetaprotein (the tumour marker) and following radiation therapy a reduction in the tumourmarker and also a complete response following radiation.just one slide summarizing our own experience and many others have also reported impressiveoutcomes following radiation of hcc with portal vein thrombus. our experience updated by bujoldin jco recently showed median survival of 17 months. the only 2 factors important foror predictive for survival in multivariate analysis were invasion into the portal vein,a known strong prognostic factor but also

the trial the patient was treated on. thefirst 52 patients did far worse than the last 52 patients that were treated. you can thatthe size of the effect is as strong as portal vein thrombus. so, why could this be? somepossible reasons for worst outcomes in the early years, hopefully we have improved ourpatient selection along the way but we actually still treat pretty large, pretty aggressivetumours, pretty unstaged tumours in our clinic today. we have definitely improved imagingat baseline, for diagnosis perhaps that has led to improved contouring and less miss fromour contouring. there are advances in image guidance i have talked about. we started withpoor man's igrt and now use state of the art volumetric 3d and 4d cone beam ct. we havehad advances in planning and i think in general

there really is a learning curve for liverradiotherapy that is something that hopefully we can shorten as other centres start up programsas we share the tips that we have come across from our own experience.so, from this experience and promising results using radiation therapy for hcc with portalvein thrombus, this has led to the phase 3 study through the rtog, rtog 1112 that israndomizing patients with locally advance hepatocellular carcinoma unsuitable or refractoryto tace so majority will have portal vein vascular invasion to standard care which remainssorafenib vs. 5 fraction sbrt followed by sorafenib. the primary endpoint is survival.we need international collaboration to make this happen; i think this is an importanttrial. it does require modern igrt credentialing

and real time qa but i think it is worth theeffort to develop a system, ensure patients can be treated safe and then we can show themaximal impact of our therapy radiation on hcc patients.in conclusion there is a strong motivation for igrt for daily liver cancer sbrt. thereare many igrt solutions but for all of them there are some common themes and perhaps themost important is that a soft tissue surrogate for the liver tumour itself is required anddaily igrt before every radiation fraction is needed. igrt increases the chance of uncomplicatedsustained local control and with modern igrt techniques there is truly less need for "imagining".now that does not mean that we shouldn't be speaking and collaborating with our radiologycolleagues. i think that is a crucial part

to a successful liver sbrt program.