Predict tool - is it ever used to agree treatment with onco?

I’m a bit confused by the Predict tool and would welcome being set straight.  

First off, I know the Predict tool is based on data and I also know that we are not statistics and we have no idea on which side of the line we as individuals will land in reality, however I have seen a lot of posts on here in which people refer to their Predict score so I assume it is being used in discussions with oncologists. 

I had a very basic conversation with my oncologist regarding treatment as, I felt, the rads and Letrozole for 5 years was a standard package for her to dish out.  I had no idea that the Predict tool existed until I saw it on here much later, so I downloaded it and had a look.  

I filled it in with the info I know and guessed at the stuff I don’t know (basically the grade and the Ki-67) and the gain on 5yrs of endocrine treatment was between 0.3% and 0.9% so there must be more to it than that otherwise why would someone like me bother?

I see the data is based on women who took Tamoxifen, who presumably are going to be younger than people who are on AIs - are we comparable?

It doesn’t include your Allred score, so would being 6/8 be a different outcome to being 8/8?

After each result it says “If death from BC is excluded, (eg) 83% survive and 17% would die of other causes”.  So if there were 200 people and 100 died of breast cancer, 83 people survive and 17 die of other causes, it still means 100 people die of BC.  Surely the point is to know how many die from BC or am I just reading it wrong? 

Has anyone actually used this tool in a conversation with their onco and made actual treatment decisions on their results? Sorry for being intensely thick if I have got hold of the wrong end of the stick on this. 

morning tigress, 

I have looked at the predict tool , you really need to know your grade to get a more accurate score . The way I read it is for example - it tells you, you have a 90% of survival at 10 years with surgery alone . If you add hormone therapy and bone stengheners this may increase to 92 %. Of these remaining 8%, 6 will die of something else in 10 years meaning you have 2% risk of dying from breast cancer in 10 years? ( These are just numbers as an example ) and of course all our cancers are unique and our lifestyles are different too , 

i will be starting radiotherapy next too . Best wishes x

I am trying to make head or tail of it myself. Last time I had breast cancer I worked out the main treatment that I had i.e. Surgery and radiotherapy was the biggest life saver. Taking tamoxifen for five years would have had a marginal effect. So I decided not to take tamoxifen for more than two months I think it was. That was about May 2004. I got new breast cancer diagnosis in June 2022. With a new primary. I am still alive. But the Predict model does seem to look fairly short term. Most of these models assume you are cured once you live for maybe ten years. At 48 that’s not that great. At 78 or 88 it might be fine…


I am not sure if you can put links on this site but let’s try 

How Predict works

The estimates that Predict produces are based on scientific studies that have been conducted into how effective breast cancer treatments are. We know from studies involving many thousands of women that the benefit from treatment is affected by the size and type of the cancer at diagnosis, whether the cancer has spread to involve lymph nodes, and whether or not the cancer expresses markers such as the oestrogen receptor (ER), HER2 and KI67. By analysing the results of these studies, statisticians are able to say how these aspects of the cancer are likely to affect average survival and how much benefit might be gained on average from different treatment options.

Predict has been tested to make sure that the estimates it produces are as accurate as they can be given current knowledge. Predict was originally developed using data from over 5000 women with breast cancer. Its predictions were then tested on data from another 23,000 women from around the world to make sure that they gave as good an estimate as possible.

Although Predict produces good estimates, it cannot say whether an individual patient will survive their cancer or not. It can only provide the average survival rate for people in the past of a similar age and with similar cancers.

The technical section has more detail on how Predict was developed and tested.

It does suggest using the tool in conjunction with your doctors. I am going to have a chat with my oncologist on 15 February over the phone, I wonder how I will transmit my scores. I know the grade of my cancer. It’s grade 2. I know the size of my tumour 2 cm (the most common size apparently). It’s spread to one lymph node although my pathology report goes on about the lymph node having metastasised due to the amount of breast cancer in it. It looked like a second breast cancer tumour in fact. So I think it may be difficult to apply the tool in some instances e.g. mine where I have intramammary lymph nodes. They aren’t very common unlike my bog standard tumour size and average grade.

A computer model which is programmed with information is likely to be less subject to bias than something which involves human intervention and judgement. So I am willing to give it close consideration.

A good predictive model also needs to be continually updated with new data. I have looked and have seen that the Predict model has been updated with a lot of new data over the years to  ensure it is still reliable.

I just had another look at the predict tool and thought it might be worth looking at. The first time I did it, I think I put in 2 for the size of the tumour in cm, but it’s in millimetres so I should have put in 20.

It did look as if there are quite significant benefits in taking letrozole and biphosphonates for post menopausal women of my age i.e. 67, compared with surgery alone.

On the other hand the quality of the life you live may not be as good. Unfortunately being dead doesn’t give you a great quality of life either.

The juries out - that’s my latest prediction.


I think it depends on working out your life expectancy with or without  breast cancer. If you are 67 as I am, and I am looking at whether I will be alive in 15 years time (the longest time interval Predict has) I will be 82 then if I survive that long.

But I could get all sorts of other conditions that could kill me off before 82 as it’s quite common for people to die aged over 70. The old idea was we live for three score years and ten, that is 70. If that is the case, I am near the time I should pop my clogs! From breast cancer or something else.


Interesting research paper on this subject, worth a look re Predict and Adjuvant Online tools

the conclusion seems to be the older you are the less you can rely on Predict to work out  your long term life expectancy if taking adjuvant therapy after breast cancer.

Please read the article to see if I have interpreted the conclusions correctly


I saw an oncologist after my lumpectomy and he said radiotherapy then 5 years letrozole. I am 74 and my tumour was 11mm grade 2. I had the radiotherapy but felt I was not advised about the radiation rash after 3 weeks. Apparently many people my age avoid radiation but I was not offered this. I was not happy with letrozole due to my osteoarthritis and saw a second oncologist, a female this time. She looked at the scores and predicted that if I took no hormone pills the risk of recurrence was low. I agreed to do 6 months to reduce oestrogen and then to stop.
It is worth having another opinion if you are unsure but ultimately the patient has to make the decisions.

Hi. Yes Oncology used it with me to help me decide whether to have further treatment after surgery. (I have had a left side mastectomy with implant reconstruction, full lymph node removal and my full nipple removed (3 small grade 1s and a minor stage 2 on size). 1 out of 6 lymph nodes had cancer in them. No evidence of spread beyond this so only a local recurrence. I have posted under ‘The Right to Refuse Tamoxifen’ on here separately).

I have not agreed to have chemo or to take the hormone blocker tablets as they came out as 1.6% and 1% benefit, and I need to be able to work and pay my bills.

The decision whether to have treatment or not is entirely yours - please see Do I have the right to refuse treatment? - NHS The key part of the NHS link on the web is ‘your decision must not be due to pressure by healthcare professionals, friends or family.’

I’m glad I read this post! I have been gaily saying theres very little benefit to hormone therapy for me acording to Predict, but I have been putting my tumour in as 5cm when it was in fact 52mm! Changes the story significantly…

Hmm so 5cm and 50 mm are exactly the same - so it only accepts mm ?

No I was putting 5 in not realising the settings were mm so it was reading it as 5mm when it should have been 52.

52mm is 5.2cm which is not that much bigger. I like many other women made the mistake of putting in 5.2 when I should have put 52 as the Predict model asks you to put the size of lump in mm (millimetres) not cm (centimetres)!

I think it’s all very confusing for the newly diagnosed breast cancer patient, especially if they decide to read the reports they get from pathologists. They are professionals and use the weird terminology of such people. I am fairly sure that oncologists don’t want people to read their reports as they may find out how bad their cancer is. In fact they would probably prefer people only to read the reports if they have a low grade cancer which has a lower chance of spreading or causing ill health.

I don’t think they want people to ask too many questions as they have hundreds of patients and haven’t got much time to see each person with a cancer diagnosis. Covid caused a backlog too.

One of them once rang me at home and in passing told me how awful it was as there are so many women she and other oncologists had to see. I think that apart from covid blocking things, this could be due to the increase in women aged 65 and over due to the baby boom of the 1950s and 60s.

I had my radiotherapy planning appointment yesterday. We were discussing the side effects of the treatment and my husband mentioned that no one has ever really discussed the exact benefits with us. Then the Predict tool was mentioned. The nurse asked if I had looked at my results. I haven’t partly because it is difficult to unknow something once you have read it. Can it be used to see the risks of having versus not having the different treatments?

Hi @scientistamafier if you haven’t looked at it (and there’s no reason that you should if you might be uncomfortable with the results) I’ll try to give you an idea.

You enter some basic data regarding your age, tumour histology and potential treatment paths. It then comes back with a statistical prediction of your survival over 5, 10 or 15 years. This is all based on tens of thousands of research results. It is predominantly about survival. What it does tell you in percentage terms how many more women will survive on your treatment plan as compared to those who do not have any component part. As an example, it may come back and say after 10 years, 77% of women will survive on surgery only, 20% will die of other things and 3% will die of breast cancer. If, say, you have had 10 years of endocrine treatment, then 1% more will survival for 10 years. The suggestion therefore is, with this set of data, one in every hundred women will have greater longevity with endocrine treatment.

So, in answer to your question, it does give an indication of the benefit of each strand of treatment (excluding radiotherapy for some reason), however, this is very broad brush because there is no telling how many women comprise the comparative data set. Could be 500 or 15,000. So that 1% could be 5 or 150 people. You just don’t know. I always say, in those numbers are tens of thousands of individual stories, so the results are only slightly better than a wet finger in the air on a personal basis. I honestly don’t know if the MDT looks at Predict, from personal experience, it was never mentioned by any specialist I saw and I got given the standard NICE recommended treatment plan even though the Predict score suggested that endocrine treatment would make virtually no difference to my survival.

Perhaps you could ask the nurse for the potential percentage benefits of each line of treatment (rads excepted) without actually hearing what the absolute percentage survival figures (according to Predict) are?

1 Like

Thanks for the great explanation Tigress.

In the end curiosity got the better of me. As you say these are all averages not specifics and, since it currently excludes radiotherapy, probably not that useful any more. I’ve already had surgery, chemo and started Herceptin. The nurse did seem to think that the stats for radiotherapy would be included at some point.

Yes I think they’re trying to improve it all of the time but it will be difficult if studies haven’t collected the appropriate data over a significant period of time. It was devised at Cambridge University, which is my old alma mater, so I have confidence in the concept but am a little more sceptical about how accurately my results apply to me. Talking of data, I’ve just realised that the percentages in my example don’t add up :woman_facepalming:t3:. How embarrassing when responding to a scientist! Good luck with the rads @scientistamafier, I hope the sessions fly by.

PS they add up now!

Predict was done for me and I thought that radiotherapy was included in it as was hormone therapy - which I eventually abandoned. Hormone therapy on top of surgery and radiotherapy only gave a improvement 0.8 - 1.3 % up to 15 years for me . I can’t remember the improvement radiotherapy made on top of surgery but I was told that it cuts local recurrence by up to 50% and continues working in the body for up to 6 years so I decided to have it. Xx

No, Predict doesn’t include radiotherapy. I have heard of people being given a percentage benefit of rads but don’t know where it comes from. Maybe it’s a blanket percentage for everyone :woman_shrugging:. I was never given any statistics or consulted about anything, just told you’ll be starting radiotherapy on this date and Letrozole tomorrow. After a year I asked to go through Predict with the BC nurse and she was quite sanguine about me giving it up if I wanted. I’m on a break at the moment and may make it permanent depending on how a few tests go.

It’s funny isn’t it I thought I had taken it all in at the Once appointment but obviously not. I’m pretty sure about the 50% reduction in local occurrence from radiotherapy though but it may have been the radiotherapy team that told me that and yes it may be a blanket percentage perhaps depending on the surgery and having achieved clear margins . Xx