Every week a patient asks me some version of the same question: “My last dentist said I had a cavity, so why aren’t you filling it?” Sometimes the reverse: “Why are you recommending a crown when it doesn’t even hurt?” Both questions come from the same misunderstanding, which is that a cavity is a single thing with a single answer. It isn’t. After more than thirty years in practice, most of them spent repairing the consequences of decay that was treated too late or too aggressively, I want to walk you through how I actually decide what to do with a cavity, what the newest diagnostic technology has changed about that decision, and why the goal is never simply to fill a hole.
Tooth decay remains the most common chronic disease in the country. According to the CDC, one in five adults between 20 and 64 has at least one untreated cavity, and national survey data shows the rate stays near that level well into the retirement years. Researchers at the University of Iowa put it plainly in a report this spring: caries is still the most prevalent disease in the world, and the profession is only now moving away from a model that waits for damage and then repairs it.
That shift matters more for adults than most people realize. The cavities I treat are rarely in teenagers. They are in patients over fifty, on root surfaces exposed by receding gums, and underneath fillings and crowns placed decades ago. Those cavities behave differently, and they demand a more careful decision than “drill it.”
A cavity is a process, not an event. Bacteria in plaque turn sugar into acid, the acid dissolves minerals out of enamel, and saliva spends the next hour or two putting them back. When acid attacks outpace repair, the enamel weakens beneath an intact surface. That is the white-spot stage, and it is the only stage where the tooth can recover on its own. I covered the biology of that recovery in Will Teeth Repair Themselves Naturally?, so I will not repeat it here.
What matters for treatment decisions is what happens next. The lesion moves through the enamel, reaches the dentin underneath, and then progresses toward the pulp where the nerve lives. At each step the surface may or may not have actually broken open. A lesion that has reached dentin but still has an unbroken surface is called noncavitated. One with a physical hole is cavitated. That single distinction drives more of my decisions than any other.
When I see a suspicious area on a radiograph, I am asking myself five questions before I say anything to the patient.
How deep is it? The American Dental Association’s guideline on nonrestorative treatment states that a lesion between teeth that appears confined to the enamel or the outer third of the dentin is most likely noncavitated, and that the dentist should prioritize nonrestorative options over a filling. In practice, that means a lesion at that depth in my office gets fluoride varnish, resin infiltration, or a sealant, and a follow-up radiograph. It does not get drilled.
Is it active? Two lesions can look identical on an image and be entirely different diseases. One is a scar from a period of poor hygiene ten years ago that has since hardened and stopped. The other is soft, chalky, and advancing. I check the surface, I look at the patient’s history, and when I have prior radiographs I compare them directly. A lesion that has not changed in three years is not a lesion I am eager to open.
What is this patient’s risk? A 35-year-old with excellent hygiene, normal saliva, and no other decay can safely watch a shallow lesion for a year. A 68-year-old on three medications that dry the mouth, with recession and two new root cavities since last spring, cannot. Same radiograph, different plan. This is why our hygiene visits include salivary diagnostics and a periodontal risk assessment rather than a quick polish; the decision on any single tooth depends on the state of the whole mouth.
Where is it? A lesion on the chewing surface of a molar, a lesion between two teeth, and a lesion at the gumline on an exposed root are three different problems. Root surfaces have no enamel, only softer cementum and dentin, and decay there moves quickly. I am far more conservative with a shallow enamel lesion than with a root lesion of the same apparent size.
What is next to it? If the suspicious area sits beside a twenty-year-old crown margin or under the edge of an old filling, the calculus changes. Decay that has crept beneath an existing restoration cannot be arrested with varnish, because I cannot see it and neither can the varnish reach it.
Here is the part patients rarely hear. Every filling has a lifespan. When it wears, chips, or leaks, it gets replaced with a larger one, which eventually gets replaced with a crown, which eventually may need a root canal. Dentists call this the restorative cycle, and once a tooth enters it, the tooth only gets smaller. That is why I decline to fill a shallow, inactive lesion in a low-risk patient. It is not that I am being cautious with your time or money, although I am. It is that the first filling on a tooth is the most consequential one, and I want to be certain it is necessary before I place it.
When do I fill without hesitation? When the surface is cavitated. When the lesion has moved past the outer third of the dentin on serial radiographs. When the patient has symptoms. When it is under or adjacent to an existing restoration. And when the patient’s risk is high enough that waiting is a bet I am unwilling to make with their tooth.
Once I have decided a tooth needs a restoration, the second decision is how much to take out, and the answer has changed since I trained. The ADA’s 2023 restorative guideline recommends what is called selective carious tissue removal for most moderate and advanced cavities: the dentist removes the soft, infected dentin and stops when firm dentin is reached, rather than drilling until every trace of discoloration is gone. Leaving that inner layer protects the pulp, reduces the chance of a root canal later, and preserves structure the tooth will need for the rest of its life.
As a prosthodontist, this is the part of cavity treatment I think about most. My specialty exists to rebuild teeth, and the single biggest predictor of whether a tooth can be rebuilt well in twenty years is how much of it is still there. Every millimeter I preserve today is a millimeter I will not have to replace with porcelain or titanium later.
The most useful development in cavity care over the past few years is not a new material. It is a second set of eyes on the radiograph.
Reading a bitewing X-ray for early decay is genuinely hard. A shallow lesion between two teeth is a subtle change in grayscale that can be hidden by overlapping enamel, angulation, or the shadow of an adjacent restoration. Two competent dentists can look at the same image and disagree. Artificial intelligence trained on millions of annotated radiographs does not get tired, does not get distracted by the patient talking, and flags the same subtle change every time.
The evidence is now solid enough that the FDA has cleared several of these systems. In July 2026, VideaHealth received clearance for a caries detection algorithm that, in its clinical trial, reduced missed cavities by 43 percent among dentists using it and cut false detections by 15 percent. Pearl’s Second Opinion and Overjet’s platform received clearance earlier for cavity detection and bone-loss measurement.
At One Manhattan Dental, AI-assisted radiograph analysis is part of the comprehensive diagnostic workup for new patients. Here is how I use it, and just as important, how I do not. The software reviews the images and highlights areas it reads as decay, with an estimate of depth. I then review every one of those flags myself. Some I agree with. Some are artifacts or old, arrested lesions the algorithm cannot distinguish from active ones, because it has never met the patient. The AI makes me more thorough; it does not make the decision. It is also a remarkable tool for the “watch” side of the equation, because it lets me compare a lesion across visits with a consistency the human eye cannot match, so a lesion that is actually growing gets treated a year sooner than it might have, and one that is stable gets left alone with confidence.
For decay hiding under existing restorations, which a two-dimensional image often cannot resolve, we add 3D cone beam imaging when the case warrants it. Between the two, the era of finding out about a cavity because a tooth broke is largely over for patients who keep their appointments.
Since it is where I spend most of my clinical time, let me be specific about decay under old dental work. Every restoration has a margin where it meets natural tooth. Over years that margin can open by a fraction of a millimeter through wear, bite stress, or material fatigue. Bacteria colonize the gap, and decay develops beneath the restoration where it produces no symptoms until the tooth fractures or the nerve becomes involved.
This kind of cavity does not get the watch-and-wait treatment. Once decay is under a crown, the crown has to come off, the decay has to come out, and the tooth has to be reassessed from scratch. It is also the reason I care so much about how a restoration is made the first time. A crown fabricated in our in-house lab with a precise, sealed margin is far less likely to leak than one made from a rushed impression, and when a patient arrives with a mouthful of aging restorations failing on different timelines, the question stops being about one tooth. That is where full mouth reconstruction comes in, and it is often the more conservative choice compared with replacing crowns one at a time as each fails.
One more tool belongs in this conversation. Silver diamine fluoride is a liquid painted onto a cavity that stops decay from advancing. It has been used internationally for decades and in the United States since 2014, and this summer a University of Michigan trial published in JAMA Pediatrics gave it the large American evidence base it had lacked, arresting decay in more than half of treated teeth with reapplication every six months. That trial was in young children, but the lead investigator noted its value for older adults and patients who cannot tolerate long procedures, and the ADA’s nonrestorative guideline already includes it.
The limitation is cosmetic: it permanently stains decayed tooth structure black. For a root cavity on a back molar in a patient whose health makes a long appointment difficult, that trade-off is often worth it. On a tooth anyone can see, it is not. In my practice it is a bridge that buys time until a proper restoration, and in some patients, a way to avoid one entirely.
None of this replaces the fundamentals, which are still what I tell every patient. Brush twice a day with a fluoride toothpaste, and ask about a prescription-strength formula if your mouth is dry or you have a history of decay. Clean between your teeth daily; most of the cavities I find between teeth would have been prevented by floss. Watch how often you eat and drink sugar rather than just how much, because a sweetened coffee sipped over three hours is worse for enamel than one finished in ten minutes. Tell your dentist about every medication you take. And keep your hygiene visits, because the cavities that cost the most are found on radiographs, never in the mirror.
A good dentist is not someone who finds the most cavities. A good dentist is someone who knows which ones need treatment now, which ones can be arrested without a drill, and which ones should simply be watched, and who has the imaging to tell the difference. That judgment is what you are paying for, and it is the reason a comprehensive exam here takes an hour instead of fifteen minutes.
If it has been more than a year since anyone took a full set of radiographs, or you have been told you have a cavity and want a second opinion on whether it actually needs a filling, call us at 212-223-3632 or book a consultation online. The cheapest cavity to treat is the one caught early enough to leave alone.
Sources cited
