Should we stop using cephalometrics in orthodontic research?
Should we stop using cephalometrics in orthodontic research?
This blog is concerned with my personal opinion on the use of cephalometrics in orthodontic research. Some people may not share this view, and I hope that this is not too controversial.
All orthodontists are ingrained, from the first days of training, in the use of cephalometrics for the clinical care of our patients. It is, therefore, entirely logical that we use cephalometric measures in research. While I do not dispute their usefulness in the past, I am beginning to wonder whether they have a role to play in contemporary research.
I would like to start this discussion by considering the work of the early orthodontic researchers. Their valuable studies were retrospective cases series based upon the analysis of existing records, such as, study casts and cephalometric radiographs. This resulted in these outcomes becoming the “bedrock” of orthodontic research and we have continued to use these outcome measures in recent studies. But I do wonder if it is time to us to re-evaluate the use of the cephalometric measure? I will discuss this under several main headings.
Does cephalometrics provide us with useful research information?
While it is easy to pose this question, it is useful to imagine that we are planning a study and the cephalogram has not been invented. I suggest that we would measure outcomes that have clinical meaning and had direct relevance to our treatment. We would also hope to keep our study as simple as possible. As a result, there would be no need for large “cephalometric festivals” of many variables reported to an accuracy of 0.01mm, with a large inherent error of the method. My opinion is influenced by many years of trying to understand these tables, and concluding that many of the differences reported are neither clinically relevant nor interesting. I too am guilty of this publishing “crime” and the evidence is in this paper.
While I have yet to meet the orthodontist who reads all the data in these tables, we should also consider whether cephalometric measurements have any meaning to those who are not orthodontists. This was a lesson that I learnt when I was applying for my first major research grant. I had carefully put together a great study design (in my own mind) and took a trip down to London to discuss this with a study advisor at the Medical Research Council (a major funding body of clinical research in the UK). When I met with him we discussed my proposed cephalometric outcome measures. He thought for a moment and then said “You are going to ask us to fund a study that aims to detect a difference of 1 degree in a measure of the inside of a person’s head. You need to know that you are competing against people who want to study HIV and cancer. I think you should go away and thinks again”. So I headed back up to the North of England and re-planned the study with an outcome measure of overjet. They clearly understood this because when we submitted the study it was funded. This was a great example of needing to move away from small measurements that only we understand to carry out orthodontic research in the real world of healthcare.
How should cephalometric measurements be analysed?
I would like to go back to basic statistics and remind us all that when we set a p value at 0.05, this means that we are taking a 5% chance of an incorrect rejection of the null hypothesis, when it is correct (Type I error). In other words we have a 5% chance of detecting a difference between two interventions, when there is none. This is a termed a false positive. In many studies investigators make this this risk acceptable by only measuring a limited number of variables.
However, if we look at any paper with a “ceph fest” of tables with statistical testing across, perhaps, 50 variables, the chance of a false positive is high. (An example is here). This problem is made worse because the individual ceph measures are not independent of each other. For example, a person with a large ANB is likely to have an increased overjet etc. Therefore, if we must use cephalometrics in our studies we need to reduce the number of variables that we measure and analyse. This was the approach taken by the UNC based early Class II study.
Is there a place for cephalometrics in orthodontic research?
At this point you can see that I am coming to the conclusion that there is a limited role for cephalometrics in our research and we should consider using other outcomes that are more important to both patients and clinicians. This is particularly fresh in my mind today because I am on my way back on the train from attending a meeting of the COMET initiative. This picture is the view on a great day in the South of England. This is a world-wide organization of people developing Core Outcome measures for trials. Researchers aim to identify outcomes that are relevant to both patients and clinicians. When investigators develop a core outcome set they ask patients about the outcomes that are important to them. I am very doubtful whether any patients will include their ANB in their list!
What should we measure?
Before I consign cephalometrics to the bin/trashcan, I should put forwards some ideas on what we should measure. In the absence of an outcome set developed using COMET methods. I suggest that these could be as simple as dental and facial appearance, the occlusion (maybe PAR), a quality of life measure and the process of treatment. If we did this I am sure that orthodontic research would be more scientifically sound interesting and relevant!
Emeritus Professor of Orthodontics, University of Manchester, UK.
Doctor, Please do not waste our time by presenting Randum Nonsense.. Are you an Orthodontist or Ortho-Dentist….. What value one can get from dump Q? Sir we treat variable es you clearly have no confidense of what you present……. Stay well…. American Orthodontics Is the Global Ki8ng we Value what we serve we do not chew fat… most of the time
Sorry that you do not like the Blog. In answer to your question, I am a specialty qualified orthodontist.
Dear Al Atta,
I personally think your comments are way off the mark here – like all Rads they must be justified and expected to ADD benefit where-ever possible, so it is not just about Lat. Cephs.
I suspect as 3D CBCT is a relatively ‘new’ medium, it too will get over-used initially, but hopefully it will be quicker than Lat.Cephs to achieve a concensus of ‘prescription value’ guidelines that will emerge to benefit all.
So yes Kevin, in broad terms I agree with the key messages in this blog too.
Yours not an Orthodontic Specialist but still does care,
Tony.
Your statement, “American Orthodontics Is the Global Ki8ng we Value what we serve we do not chew fat… most of the time” is incoherent.
Dear Al Atta, I don’t think there will be any orthodontist in the world who will be unaware of Prof O’Briens name. I would like to add that cephelometrics do have a role to play in orthodontic teaching as the pre and post superimpositions give a deep insight into the treatment results. In my opinion the ceph is being over prescribed now either because of a habit of prescribing it before the treatment starts or for medico-legal purposes.
The “Cephalometric Festival ” (love the term !) is sacrosanct to a regimentally trained orthodontist, who over a period of time does tend to fall into a “what needs to be done” rut rather than “why” it should be done. I concur with your views Prof O’Brien , that quantifiable variables that affect treatment outcomes that are clearly discernible by a patient, or on facial appearance are parameters that need research and database . Cephalometrics has provided useful information in the past, but could not have that been derived by other means , is a question we need to ask ourselves . With the advent of 3D Diagnostic Methods and Customised appliances , treating to numbers or researching their relevance ….for sure needs a rethink !
Thanks for your comment. I also wonder what the use of 3D analysis will do to research and I hope that researchers are careful in the number of variables that they report. I think that the approach taken by Lucia Cevidanes and Hugo De Clerck in their Class III studies is looking sensible.
I wonder that 3D analysis will be the next protocole for research. May be by structures analysis insteat of a lot of variables and treating numbers.
One must think again in the future not in the past.
I think you are on to something. However, I still feel cephalometrics can still be useful in helping us make decisions about treatment in difficult cases. It will be interesting to see how 3D analysis evolves.
the issue is that, there are no standard or accurate measures of dental ,facial appearance or quality of life
Kevin
as usual a thought provoking blog – you have summerised what i have thought for some time, Cephs have limited relevance; they are a 2D representation of a 3D object (the head) a is centered on the midline. What are your thought on using 3D imaging, as the statistical analysis is now available to monitor and evaluate 3D change, allowing for growth and development.
Ross
Well said Prof. Kevin and Prof. Nikhilesh…I too have seen throughout my orthodontic training a lot of my fellow senior and junior ortho residents working on cephalometric studies which I hardly found clinically relevant..the only studies which stand test of time are the ones which makes sense to the physical world like growth studies, skeletal maturity indicators etc. which can be deduced and made understandable to a parent on why his/her child’s growth is in wrong direction and how can it be corrected..on the latest diagnostic technologies like CT all the studies being done in well equipped centers have CT pushed in (because it is termed as base “standard” of technology a paper should use to get published)..I would like all the trained orthodontists to ask this question from bioethics point of view..do we need CT for what can be seen by naked eyes and felt by doctors trained fingers?!
What is significant to a patient and what is significant to the specialty are two different things. Experience teaches that there is no limit to the nonsense that will be advocated by dental/orthodontic entrepreneurs. In the absence of a means of testing the most florid of claims, you have issued a license for the uncritical (or worse) to sell the unlikely to the unknowing…forever. The fact that cephalograms no longer can be taken routinely and often fuel retrospective studies, a cynic might conclude that your suggestions are designed to limit our data to a type that would favour the RCT. Granted, your measures might be useful in treatment planning (assuming that some may still want to go beyond whatever the appliance du jour happens to produce); however, they probably can’t answer our most important research questions.
Dear Dr. Kevin, first of all, I would like to congratulate you for your wonderful reflection, which I am 100% agreed. I have been working with CBCT and 3D imaging for orthodontics since 2006 and I wondering when orthodontists will realize that we are living in a “different world” in terms of what is really importante to see and detect during the diagnosis and what is important to offer to our patients in terms of quality of life improvement. Something related to “Enhancement Orthodontics” from Dr. Ackerman is the future, I think… BTW, concerning CBCT applications, what do you consider important to see and “measuring” in 3D? Best Regards, Mauricio
Loved your blog “A novel Orthodontic Paradigm” and refuse to close that window on my iPhone so I can read it occasionally
Here again in this blog you speak the wisdom that comes with experience. It would be wonderful to meet you one day. Surprised the first writer was so derogatory and had not heard of you. Possibly was not a specialist himself.
Cephalometrics is a great training tool , research tool and often but not always a great assist for treatment planning. For some it is also a great item number to add on an account. Whoops hope that was not cynicism rearing it’s head. I like all on this blog use Cephalometrics as needed but not routinely. Keep up the great work !
Michael Hyde
Australia
Yes – Specialist Orthodontist 🙂
Is it true that Stainer created his analysis based on his son’s face proportions?
I am not sure. I was always told that this was based upon the measurements derived from a beauty queen? Perhaps someone can clarify this?
Kevin,
Always a pleasure to read the blog, but posts like these go the extra mile and give food for thought.
Personally, I don’t know which seems more problematic to me: the incredible ceph-fest in orthodontic research that gives a new meaning to the term Type I error (this kind of data torture would give any methodologist the fit) or the actual need for cephalometric in every possible timepoint of orthodontic treatment. The time has come I think for the orthodontic community to seriously debate the second point. For the first point I can think of intermediate solutions like adopting a more conservative P-value (like 0.01 or 0.001) for the minimum statistical significance in the cephalometric analysis, choosing only 2-3 main ceph variables to test or, even better, use more patient-friendly or patient-centered outcomes, like you mentioned in your blog.
Spyros
The classic approach to the multiple comparison problem is to control the familywise error rate. Instead of setting the critical P-level for significance, or alpha, to 0.05, a lower alpha is used. The most common way to control the familywise error rate is with the Bonferroni correction. The significance level (alpha) for an individual test is found by dividing the familywise error rate (usually 0.05) by the number of tests. However, and importantly, the Bonferroni correction assumes that the tests are independent of each other. Kevin, as you have already alluded to, a large number of cephalometric variables are inter-related.
An alternative approach is to control the false discovery rate. This is the proportion of “discoveries” (significant results) that are actually false positives [Benjamini and Hochberg (1995).] Put the individual P-values in order, from smallest to largest. The smallest P-value has a rank of i=1, the next has i=2, etc. Then compare each individual P-value to (i/m)Q, where m is the total number of tests and Q is the chosen false discovery rate. The largest P-value that has P<(i/m)Q is significant, and all P-values smaller than it are also significant.
The goal of multiple comparisons corrections is to reduce the number of false positives. An inevitable by-product of this is that you increase the number of false negatives, where there really is an effect but you don't detect it as statistically significant.
Imaging forms only one part of the patient examination database. Rather, it is the collective of information derived from analyses of facial, skeletal and dental structures that help formulate a specific biomechanical treatment plan. Please feel free to dismiss the statistical ramblings of an occasional tooth-aligner who routinely records cephs on all patients receiving comprehensive care to not only plot and navigate the course ahead, but to keep an eye on where departed from. Whilst not every cephalometric value that my digitisation software spurts forth is directly useful, there remains a cogent few that will influence treatment planning decisions for a patients beyond whether they can be treated by a simple “non-extraction” approach.
This is an interesting discussion. Cephalometrics in the past has allowed the orthodontist to see and analyze more than before it existed. However with the advent of CBCT new standards need to be thought out. The problem with any ‘standard’, however. is that it can be accepted as gospel and therefore can have a limiting effect on treatment. Examples of this exist in our past and present, which require extractions or surgery to achieve a certain cephalometric standard. Adherence to these ‘standards’ by well meaning doctors has been a grave disservice to our patients. These doctors have been ignorant of very real side effects on the rest of the body, outside his/her view, being narrowly focused on the mouth to create straight teeth. I personably believe that we as orthodontists need to understand the effect of our treatment on the rest of the body. In my book Straight Talk about Crooked Teeth, I lay out Nine Keys to Lower Facial Harmony. These keys can be assessed with 3-D imaging along with a physical examination. These treatment objectives can be routinely achieved by the knowing specialist and will have a very positive effect on the health of the entire body. We can provide much improved treatment for this powerful specialty we love! Kent Lauson, Aurora, Colorado, USA
The MOrth exams perpetuate the need for lateral cephs. I have discussed this issue with past and present Chair of examiners. I feel many lateral cephs are taken to allow the treated case to be used in the exam. I also suspect that in a few years time our traditional methods of superimposition will be revealed as not valid, Bjorks structures and predictors of rotation will cease to be used. Research using Bolton data as control (or any other from decades ago) is also asking for trouble..
Dear Kevin,
Thank you for the blog and the reassurance that “we are not alone” in this way of thinking….
I totally agree with your comments regarding the questionable use of cephs in orthodontics – it is just so hard to strike a balance between clinical need, research need, exam need and medico-legal need!
Again I agree clinical need is fairly low but as mentioned in a previous post the MOrth exam would be much shorter without a ceph superimposition or two to “chat about”!! So not sure what the answer is??
My area of research and clinical interest is in 3D orthognathic planning (& orthodontics) and I get too lecture over the world and it still “shocks” me that some countries advocate the use of cephalometric analysis /planning for maxillary and mandibular position!! Which I think is even more “scary” compared to there orthodontic use! – Orthognathic surgery by numbers! It is very difficult to “convert” peoples way of thinking even with evidence……
Now getting back to 3D, I agree this will be the future but with regard to orthognathic treatment we need to sensible and decide which cases would benefit. Think back to the introduction of distraction, when the surgeons would distract anything that moved, but they soon realize case selection is important and sometimes conventional is better than state of the art.
3D imaging generates a massive amount of information and again I think we have not actually “cracked” a method of analysis that is valid and / or clinically understandable and useful – but I am working on it….as is Ross, Lucia etc.
In summary I agree cephs have limited value in contemporary orthodontics but assuming 3D imaging will be a substitute maybe a little premature. I think the future lies in “backward planning”, get the soft tissue in the right position (the outcome measure the patient scan see) an then work out what tooth and jaw position is required. This is complex, relying on soft tissue modeling and patient-specific 3D virtual plans. But I and a group of people all over the world are working on it – so keep your eyes open…..
Thanks again for a thought provoking blog
Take care
Sunny Khambay (Hong Kong)
to asome extent Iagree with you my prof , also for my dear orthodontist ALatt , I would like to say that even among American orthodontists there are who have an oponion which is close to DR kevin O’brien view ,one of them is dr Nanda who I heard from him directly in aconference that ” cephalometry is adyind feild” , also aI think that there are many orthodontist all around have aclose view,the question now ‘If ceph, is not an accurate diagnostic tool what is plan B’
my dear dr Obrien I’m happy to say that now I’m working on aresearch about the effectiveness of ceph as adiagnostic tood and I will be glade to send it to you after finishing it
hane anice time all
Dear Kevin,
I am a late-comer to the party here, but would like to make a few comments as a nom-specialist, with about 900 UOAs worth of ortho a year.
I would like to see your comments about the use of cephs in ordinary day-to-day orthodontic practice please. I have, personally, rarely found that they influence a treatment plan. I do not have one, but consider that if the patient is likely to need one ie surgery is a distinct possibility, then I refer to my local consultant.
Also I attended a local orthodontic group meeting in a neighbouring town about four years ago. This very well run group, headed up by the local consultants, has an accreditation scheme for its local GDPs, who can submit some cases (20 I think) and sit a viva, to achieve local recognition of competence in orthodontics. Not, of course, even vaguely a registerable qualification but useful for busy GDPs. Sadly all cases had to be attended by a ceph. And since I don`t have one I could not take part. Another GDP sitting next to me felt the same way. At the time, the hospital local to me would not just do cephs – too busy.
I`d be interested to learn what people think nowadays.
Andrew Adey, Wolverhampton
Hi Andrew and welcome to the blog, I hope that you find it useful. Over the years there has been a large amount of research on whether cephalometrics influence treatment plans and this has generally shown that it does not. While most of the samples of operators have been experienced orthodontists, but some studies have used post grad students and their decisions were not influenced. My feeling about the ceph is that when we see a patient and examine them, will taking a ceph change my decision? As a result, I take less cephs. There is certainly a need if we are trying to decide whether surgery is an option and my rough rule of thumb is that as the skeletal component of a malocclusion becomes more marked then a ceph is more likely to provide important information. But if the patient is Class I or has a mild Class II or III discrepancy, then I am happy to plan and treat without the ceph. But I do take good photographic records!
Thanks very much for your reply. This is as I thought it to be – good to hear it straight from an expert and keep up to date!
Regards
Andrew Adey
My son is currently finishing up his orthodontic treatment. The orthodontist is saying he will be finished around spring and this is November.
I allowed the orthodontist to take a cephalogram, and panoramic x-ray before treatment. Now he is saying he would like another cephalogram and I am pretty sure he said it is to determine if a lower front incisor’s root is twisted. I think I am going to say no to this because I feel afraid of too many x-rays and my son has had a few in his life. Do you feel that it will be putting the results in jeopardy or that there is another way for the orthodontist to assess the tooth? Or is the idea that he has of the root twisting kind of weird?
Your thoughts would be appreciated.
Hi Donna, thanks for the question. It is very difficult for me to comment on your question because I have not seen your son or his radiographs. But I would discuss your concerns with your orthodontist and explain that you are not keen on any further x Rays and could he give you more information on why they need more information. They should be happy to do this and certainly give you sufficient information for you take a decision.
I partially agree with your opinion, if you look at the methodology of the most used cephalometric analysis there is an evident lack of proper study design. But I think on cephalometrics like a Diagnostic test and considering that every study should be compare with a gold standard to measure the sensitivity and especificity of the test. There`s only a few studies with this type of methodology and i found that very useful, because with the R.O.C. Curves you can have the specific value of sensitivity and specificity of given cephalometric measure (You can know how accurate is an angle in classifying a class I, II or III, etc). Is evident that new technology will change ceph analysis, but when you properly use cephalometrics you can obtain some data that will help you with diagnosis.
Regards,
Pablo Gálvez
P.D.: Excellent Blog!
Hi Pablo, thanks for the comments. You have highlighted one of the major problems with cephs in research and this is that we do not have a gold standard analysis. Neither do we have a set of cephalometric records that we could use to calibrate investigators. I think that it would be incredibly useful if we could reach agreement on what measurements we should make in a ceph study. But I am not sure that this is possible because most people have their own favourite ceph analysis etc!
though late to this discussion i still appreciate the opinions of the various contributors. to prof. Brien thanks for starting this. to AL Alta there is sense in every nonsense. stupid ideas today turn great ones tomorrow.