How reliably does a radiograph detect calculus?

In two studies that checked radiographs against the extracted teeth, the radiograph's sensitivity for calculus was 44% and 50% [1, 2]. Guidance still lists calculus deposits among the things to assess on a periodontal radiograph [3], and a 2026 scoping review found a significant gap in research validating that use [4]. What a clear radiograph tells you is the part the evidence hasn't settled.

What the question actually is

Detecting calculus on a radiograph means a reader looks at the image and calls a deposit present or absent. Whether that call is right can only be judged against something else, and what that something is decides what a study can tell you.

Two studies test the radiograph directly, using the tooth itself as the reference. Radiographs were taken before extraction, and the same surfaces were examined after it.

• Buchanan and colleagues took standardized periapical radiographs of 275 proximal surfaces from 18 patients, read independently by two investigators under optimum viewing conditions. After extraction, the teeth were examined under the microscope and by planimetry on 40× linear projections [1].
• Hyer and colleagues took digital radiographs of 71 hopeless teeth from 34 patients, read by two examiners. After extraction, the teeth were stained with 1% methylene blue, each interproximal root surface was photographed, and the calculus was measured as a share of that surface [2].

That design gives two numbers, and they answer different questions. Sensitivity is the share of surfaces with calculus that the radiograph called positive. Specificity is the share of calculus-free surfaces it called negative. Keep them apart when you read a figure quoted anywhere else.

Change the reference and the result changes with it. A 2018 study of 1,091 health-checkup patients in Korea used panoramic radiographs as the reference standard. Against them, the oral examination's sensitivity for supragingival calculus was 18.3%, and subgingival calculus was found only on the radiographs [5]. A design like that can't measure what the radiograph misses, because the radiograph is the answer key.

The same goes for studies that use the radiograph as the measure. One study of 622 periodontal patients counted radiographically detectable subgingival calculus on periapical radiographs, scoring deposits of 1 mm or more [6]. In deep-learning work on bitewings, calculus is whatever the readers labelled: in one dataset of 4,837 routine bitewings labelled by two calibrated dentists, the prevalence of calculus on image level was 36.1% [7]. None of these is an accuracy figure for the radiograph.

What the evidence shows

Both of these studies found that the radiograph missed a large share of the surfaces carrying calculus.

In the 1987 study, radiographic deposits were detected on only 44% of the proximal surfaces that showed calculus under the microscope [1]. In the 2020 study, the average sensitivity of digital radiography was 50% [2]. That's where the two agree: about half or fewer.

Read the other figures for what they are. The 7.5% in the 1987 study is its false-positive rate, reported alongside high specificity; the abstract gives no predictive values [1]. The 82.2% in the 2020 study is specificity, not a detection rate [2].

The 2020 predictive values come from hopeless teeth. In that sample, a positive call was right 94% of the time and a negative call only 23.2% of the time [2]. Predictive values depend on how common calculus is in the teeth being examined, so these two figures belong to hopeless teeth, not to a routine recall patient.

The amount of calculus mattered in both studies. In the 1987 study, detection was influenced by the thickness of the calculus, the share of root surface it occupied, and tooth type [1]. In the 2020 study, sensitivity rose as the area and size of deposits increased, and coverage of more than 30% of the interproximal root surface was associated with better detection [2].

Image processing didn't close the gap. In the 2020 study, auto-contrast, emboss, invert and sharpen enhancements were not statistically superior to the original image [2].

The 1987 authors put their conclusion bluntly: present radiographic techniques were "not appropriate for detecting calculus on root surfaces" [1].

Where it's unsettled

The evidence base is thin. A 2026 scoping review of calculus detection and removal after non-surgical periodontal therapy, which excluded in-vitro studies, implant studies and studies with a sample size under 30, included 59 studies. Only two of them explicitly mentioned using radiographs to detect calculus, and the authors describe a significant gap in the literature validating radiographic assessment of subgingival calculus [4].

Both sides of the split are on record. The 1987 authors concluded that "present radiographic techniques" weren't appropriate for detecting calculus on root surfaces [1]. The 2026 reviewers concluded that radiographs, combined with traditional detection methods, "may prove beneficial" after non-surgical therapy, while noting how little research validates it [4]. The same review found that no single method of calculus detection and removal was universally effective [4].

Why do the figures differ between studies?

The two accuracy studies differ on nearly everything that could move the number [1, 2]:

• Receptor: conventional radiographs in 1987, digital in 2020.
• Reference: microscopy and planimetry, against methylene-blue staining and photographs.
• What was scored: proximal surfaces, against the largest interproximal deposit on each tooth.
• Sample: 275 proximal surfaces from 18 patients, against 71 hopeless teeth from 34 patients.
• Date: 1987, against 2020.

Because they differ on all of these at once, the gap between 44% and 50% can't be pinned on any one of them.

Amount of calculus is the one factor both studies measured, and both found it mattered [1, 2]. For a sense of what is usually there: on 90 extracted teeth from 29 patients with chronic periodontitis, inspected under a magnifier, most deposits on the proximal root surfaces were thin, smooth veneers [8]. None of the studies here tested how often deposits like those show on a radiograph.

Site made less difference than you might expect in the one study that looked. In the 1987 study, detection didn't depend on which proximal surface, which arch, attachment loss or probing depth; tooth type did matter [1].

View is open. The 1987 study used periapical radiographs [1]; the 2020 abstract says only "digital radiographs" [2]. No study found for this article compares bitewing and periapical views for calculus against the teeth themselves. The one bitewing calculus study cited here is a deep-learning paper scored against dentists' labels. It describes calculus on bitewings as small irregular objects with often blurred boundaries, hard to label accurately, and notes that the main indication for bitewings is caries diagnosis [7].

What this means at the chairside

A clear radiograph doesn't rule calculus out. With sensitivity at 44% and 50% in the studies that checked against extracted teeth, a surface that looks clean on the image may still carry calculus [1, 2].

When you do see a deposit, record it. SDCEP guidance includes the presence of subgingival calculus in the radiographic periodontal report, and the British Society of Periodontology lists it among the things radiographs should be assessed for [9, 10]. In hopeless teeth, positive calls were right 94% of the time [2].

The radiograph supplements the clinical examination; it doesn't replace it. SDCEP puts the clinical examination and full-mouth periodontal assessment first, supplemented by radiographs where clinically indicated, and says radiographs let you assess "sometimes subgingival calculus" [9]. The ADA/FDA selection criteria list calculus deposits among the structures to assess on radiographs of periodontal patients, alongside the level of supporting alveolar bone, the condition of the interproximal bony crest, the length and shape of roots, and bone loss in furcations. They describe recall radiographs as selected intraoral views used to verify clinical findings on a patient-by-patient basis [3]. The BSP notes that bitewings, taken routinely for caries, may also give early warning of subgingival calculus and of poorly contoured restorations [10].

The clinical check misses calculus too. In one study of 101 teeth extracted after scaling and root planing, 57% of 476 surfaces still carried calculus under the stereomicroscope, and the mean percent calculus per surface area was 3.1% (0 to 31.9%). Clinicians had scored 77.4% of those surfaces as free of calculus; 11.8% of the surfaces microscopically free of calculus had been scored as carrying it, and agreement between and within examiners was low [11]. Across 4,160 surfaces scored after student scaling, agreement between clinical examiners was a kappa of 0.33, and on manikin teeth with simulated calculus, raters found 49% of the deposits [12]. In one study of seven patients, none of probing depth, bleeding on probing or attachment level provided both high predictability and high sensitivity for residual calculus [13].

On this evidence, neither the radiograph nor the clinical check clears a root surface on its own. The guidance order is the clinical examination first, with radiographs to supplement it and verify its findings [3, 9].

You've seen what a radiograph can and can't show of calculus. Try reading real bitewings.