Radiotherapy 102

Judging from the last article, radiotherapy sounds really simple. By not mentioning some of the practical issues and hiding others in theoretical constructs (eg tumour region) what I did is oversell radiotherapy without actually lying. Ah, the magic of language. As always in life, the actual difficulties in radiotherapy are in the subtle day to day details. I will pick one issue as to list all would require a thick textbook (and the dear folk at Amazon will sell you any number of these). So the selected exemplar radiotherapy issue is: What do you actually want to irradiate?

Very rarely can the full extent of tumours be clearly seen on a patient's surface. This is, for the reasons explained in 101, especially true for patients advised to have radiotherapy. So whilst it is all very nice being able to conform your delivered radiation beam to a very complex geometric shape, by itself,  this is not actually very useful. All your delivered radiation being within a millimeter of some shape is only useful if the tumourous tissues actually are in this shape. Unfortunately, it is rare we are certain about the exact shape and extent of the tumour.

"Yes, but we can use volumetric CT scans accurate to 200 microns" comes the physics student reply. True, but is a tumourous tissue distinct to normal tissue on CT scans (consider what a cancer cell actually is when thinking about this)? Often tumour and healthy tissue is enitrely indistinguishable on CT, especially given there might be local inflammation or compressed normal tissue due to the presence of the tumour. When we need to take a section of tissue and decide if it's cancerous CT is never used. Mostly, only chemical stains and analysis under a microscope can identify tumour tissue with any confidence. Even then,  it is very hard and close to impossible if you didnt have "normal" images/parts to compare with.

When doctors mark where they think the tumour is on the CT data volume, a process called outlining, it is often done on the basis of what feels or looks wrong. For some tumours it is educated guessing at best.

Hence, as you might expect, the variability in outlining results is huge. A small easily treated tumour that is badly outlined as a larger tumour will result in much more radiation being requested and ultimately more side effects occurring. Bad outlining might even be the basis for declaring a tumour unsuitable for radiotherapy.

Physicists and mathematicians doing oncology research tend to ignore all this and take the doctor outlined volume as the gospel truth (partly as we prefer pleasant abstract geometric problems over messy badly formed real life biological ones1). Countless radiotherapy articles, including some by me, have been written on complex techniques and expensive technology to conform the delivered radiation more exquisitely to the outlined volumes. Even ignoring the, often unassessed, difficulty and cost in properly implementing these advanced methods, whether these techniques and technologies would be of any advantage to actual patients is actually quite arguable (remember the amount of delivered radiation has generally been tweaked by clinical experience with more standard treatment set-ups).

Surprisingly, well defined practices, such as adding safety margins after outlining, and the averaging of errors, means research, including mine, does cause incremental improvements over time. However, it is not certain this averaging out by experience would still hold if you change the process significantly (Proton therapy I am looking at you).

Please dont take this as a message radiotherapy does not work. It is stunning successful and continues to improve even with its meagre research budget2. This old simple idea and its regimented practice cures thousands in the UK every month and for that we should all be very thankful.


  1. Over simplified views of other fields is traditionally associated with physicists for a good reason 
  2. Compared to the generally much less successful field of chemotherapy. Whilst it is not a competition (both methods are often used together), nevertheless, if even a small percentage of the chemo research fund was diverted to radiotherapy many people would get to live longer.