It is difficult to determine the amount of time that may be required for a periapical bone lesion to heal. A majority of root canal treated teeth with initial bone destruction show signs of healing within 1 year [24]. However, in individual cases, the healing process can take a long time [22]. In a study by Molven et al. [23], it was reported that some cases had required more than 25 years to completely heal. The finding that there is no absolute time limit for when healing may eventually be diagnosed can also be deduced from epidemiological studies [16].

The diagnosis of periapical tissues based on intraoral radiographs is subject to considerable interobserver and intraobserver variations [12].

There are also uncertainties regarding the validity of the radiographic examination. Only a limited number of studies have compared the histological diagnosis in root-filled teeth to radiographic signs of pathology [12, 25]. In these studies, false-positive findings (i.e., radiographic findings that indicate apical periodontitis while histological examination does not) are rare. The number of false-negative findings (i.e., radiographic evaluation indicates no apical periodontitis while histological examination gives evidence for inflammatory lesions) varies between the studies. However, from experimental studies, it is well known that bone destruction and consequently apical periodontitis may be present without radiographic signs visible in intraoral radiographs.

The advent of cone beam computed tomography (CBCT) has attracted much attention in endodontics in recent years [12]. In vitro studies on skeletal material point to that the method has higher sensitivity and specificity than intraoral periapical radiography. The higher sensitivity of CBCT is confirmed in clinical studies. CBCT provides a three-dimensional image of the area of interest, an advantage when assessing the condition of multirooted teeth. As a result, the reliability of results of endodontic treatment in follow-up using conventional intraoral radiographic technique has been questioned [26]. It has been suggested that CBCT should be used in future clinical studies, because conventional radiography systematically underestimates the number of teeth with osteolytic lesions. In this respect, long-term studies are required to investigate if healing of periapical bone destruction may take longer than previously assumed. Also, there is not enough scientific evidence to tell whether lesions on CBCT images provide accurate clues as to the histological diagnosis present. So far, there are also disadvantages of CBCT, such as greater cost and a potentially higher radiation dose. Up to date, there is no evidence that suggests that individuals with subjectively and clinically asymptomatic root canal treated teeth with normal appearance of surrounding bony structures in an intraoral radiographic examination would benefit from further evaluation by a CBCT scan.

Besides the time aspect of periapical bone lesions, there is also a problem with determination of what should be considered healing sufficient to constitute a “successful” endodontic treatment. As a consequence, the question of what establishes a “failure” and hence an indication for retreatment is far from unmistakable. According to the system launched by Strindberg [22], the only satisfactory post-treatment situation, after a predetermined healing period, combines a symptom-free patient with a normal periradicular situation. Only cases fulfilling these criteria were classified as “success,” and all others as “failures.” In academic environments and in clinical research, this strict criteria set by Strindberg in 1956 [22] has had a strong position.

The periapical index (PAI) scoring system was presented by Ørstavik et al. [27]. The PAI provides an ordinal scale of five scores ranging from “healthy” to “severe periodontitis with exacerbating features” and is based on reference radiographs with verified histological diagnoses originally published by Brynolf [25]. The results from Brynolf’s work indicated that by using radiographs, it was possible to differentiate between normal states and inflammation of varying severity. However, the studies were based on a restricted biopsy material to upper anterior teeth. Among researchers, the PAI is well established, and it has been used in both clinical trials and epidemiological surveys. Researchers often transpose the PAI scoring system to the terms of Strindberg system by dichotomizing scores 1 and 2 to “success” and scores 3, 4, and 5 into “failure.” However, the “cutoff” line is arbitrary.

The Strindberg system, with its originally dichotomizing structure into “success” and “failure” has achieved status as a normative guide to action in clinical contexts. Consequently, when a new or persistent periapical lesion is diagnosed in a root-filled tooth, failure is at hand and retreatment (or extraction) is indicated.

However, as early as 1966, Bender and colleagues [28] suggested that arrested bone destruction in combination with an asymptomatic patient should be a sufficient condition for classifying a root canal treatment as an endodontic success. More recently, Friedman and Mor [29] as well as Wu et al. [30] have suggested similar less strict classifications of the outcome of root canal treatment.

The presence of subjective or clinical signs of failed root canal treatment is only occasionally reported in published follow-ups and epidemiological studies. The results are measured thus exclusively through an analysis of X-rays. In epidemiological cross-sectional studies, the frequency of periapical radiolucencies in root-filled teeth varies. In a systematic review of the matter, Pak et al. [31] included 33 studies from around the world with frequencies of failed cases varying from 12 to 72 %. The weighted average of periapical radiolucencies in the 28,881 endodontically treated teeth included was 36 %. The high frequency of root-filled teeth with periapical bone destructions seems to persist regardless of the fact that technical quality has improved over time [31]. Yet, cross-sectional studies cannot distinguish between cases that will finally heal and osteolytic lesions that will persist. On the other hand, longitudinal studies have shown that root-filled teeth without periapical radiolucent areas may develop visible lesions over time [16].

To retreat or not to retreat “an endodontic failure” is the issue. It has been argued that both modern medicine and dentistry face fundamental ethical problems if too rigorous and consistent concepts of disease prevail. The discussion about different concepts of disease goes back to ancient philosophy and has bewildered and occupied philosophers ever since. In this book about molar endodontics, we can only hint at the fundamental issues. For further reading, the interested reader should seek in books on philosophy of medicine [35].

Two fundamentally different concepts of disease can traditionally be recognized.

These theories address different aspects and pose different challenges to medicine and dentistry as a whole and therefore also to endodontics. But the two predominant concepts have been challenged for several reasons. For example, they neither separately nor together fully acknowledge all important perspectives on human disorders. An alternative approach is to apply the “triad of disease, illness, and sickness” [36] (Box 8.3). Despite criticism, the triad is widely used and discussed. The definition of the triad’s different components is by no way clear-cut. The triad and its implications on dentistry were elaborated by Hofmann and Eriksen [37]. Kvist et al. [38] made cautious and initial attempts to apply the theory to the problem of asymptomatic root-filled teeth with radiographic signs of apical periodontitis.

The three approaches to disease do not replace but complement each other. It is also the case that they are strongly intertwined. However, using the above matrix of “disease,” “illness,” and “sickness” makes it easier to understand the variation in clinical decision-making regarding root-filled teeth with persistent apical lesions, both when it is tested in different setups by researchers as well as in the clinical situation with an individual patient.