Surgical guide accuracy - intra-oral or model scan?

Would you prefer using intraoral scan OR model scan (using dental models) to be used for surgical guides and why do you prefer one not the other?

I have seen a few cases where the surgical guides cannot fit well. They are designed based on a model scan. The lab was using some other planning system, which did not have a function helping to identify the problem. The doctor was not happy, but lab didn't know why. There is no doubt a model scan has too many steps, and each will introduce errors. 

On the other hand, not all intra-oral scanners are made the same. I have seen many flaws on the models, such as overlapping, holes, the surface sticking out, etc, which may make the surgical guide design fail. This is why we spent a lot of development effort to streamline the surgical guide design.

Ideally, if you have a perfect stone model, I would think a desktop scanner is more accurate. The reality is unfortunately the other way around. With my experience, model scans tend to have "systematic" or "global" errors since the model or impression might have global deformation. On the other hand, intra-oral scans tend to have "local" errors due the scanning process, data acquisition, etc. If you would like to find out how to deal with those errors in treatment planning and surgical guides, please let me know. I can explain a bit more how GuideMia works. 

Since CT was discussed in previous comments, I would like to talk about accuracy in a more complete way. When we talk about surgical guide accuracy, there are two factors: how the guide will fit onto the anatomy, and how the treatment plan will be transferred onto the model from the CT data. Having a guide that just fits properly does not mean you have an accurate guide. 

It is obvious that optical scan is more accurate than CT scan. Many people are using this practice now, except for the cases in which radiographic guides are required. To me, a bigger source of error is still the CT scan. Treatment plan is transferred to the model by registering the model and CT scan. With our experience, a well prepared case will have about 0.05mm accuracy, some higher, some lower. If this number goes to 0.2 level, there might be something wrong with the model, or CT. 

Interestingly, the deviation between optical scan and CT scan is not very sensitive to the slice thickness of CT scan. On the other hand, the registration between two CT scans for radiographic guides will be more dependent on the CT slice thickness. 

There are clinical cases, where the registration between model and CT is not quite feasible, so there is an approach to register patient CT, radiographic guide CT and then optical scan to guide CT. I like Larry's comments about how many steps it takes to make a model scan. Same thing here. The more datasets that are involved, the bigger chance of error. 

A little bit more: Medical CT or CBCT?

Medical CT has much better "contrast resolution" (or other words gray levels), while CBCT has better spacial resolution (smaller slice thickness). This is why medical CT looks so much better than CBCT on screen, altough geometric details may be lost. On the other hand, CBCT has less gray levels. Sometimes - I don't know whether it's because of machines or becuase of settings, but CBCT datasets pretty much give you something that has two levels: bone and none-bone, which is sad. In this case the chance to misclassify a voxel is high, and then you can end up with a less accurate model even though its spatial resolution is high.This might be a bit off the topic because the topic is about intra-oral scan. With treatment planning software, we have some best practice guide depending on the data is from CT or CBCT, in order to have a better registration with optical scan. Best,Frank Gao

Contact us

5030 Katella Ave, Ste 209
Los Alamitos, CA 90720

Latest Tweet

About two years ago, our client posted the picture. Now we are seeing this approach gets some more attention. Great! https://t.co/xBwxSf3tiZ

Our Services

Social links