The cross-sectional study was conducted in Bharati Vidyapeeth Dental College and Hospital, Sangli, Maharashtra, India, in the month of January 2018 and the time required was one and half month to complete the study. Ethical clearance was obtained from Institutional Ethical Committee (IEC approval number: BVDUMC&H/SANGLI/IEC/285/18) of Bharati Vidyapeeth (Deemed to be University) Dental College and Hospital Sangli, Maharashtra, India. The sample size was calculated before the study and it was estimated that a sample size of eight dental study models and digital models was needed to obtain a statistical power of 95% [11].

Inclusion and Exclusion criteria: The study sample comprised of eight rubber based impressions of maxillary arch based on inclusion criteria of intact dentition with complete set of teeth with erupted second molars. Exclusion criteria were presence of any prosthesis, restorations, fractured teeth, presence of any developmental anomalies or gingival pathologies.

These maxillary arch impressions were scanned same day the impression was made with LED scanner (Medit Identica blue, Seoul, Korea) having twin camera with 1.3 mega pixel. Once the scan was complete, the raw images were converted to STL files to generate 3D dimensional models. With the help of the commercial software (Ansys Inc. version 19, Canonsburg, Pennsylvania, US) curvilinear measurements were obtained. The distance over a curved surface with shortest path of the curve is measured by the ruler tool.

Dental stone was poured in impressions and study models were prepared. Measurements on dental study models was done with brass wire which was marked from the beginning of the landmark ie., cusp tip to the gingival margin along the long axis of tooth for central incisor and canine and interdental contact points for 2^nd premolar and carefully wrapped along the curve being measured. Once second landmark is reached another mark was placed. The brass wire was then straightened and the distance was measured with the help of digital caliper (aerospace).

Measurements were performed in different aspects of the maxillary arch in various directions. The following curvilinear measurements were made: the labial and palatal surface of central incisors on left side along the long axis of the tooth from the cusp tip to the gingival margin [Table/Fig-1]; the labial and palatal surface of canines on left side along the long axis of the tooth from the cusp tip to the gingival margin [Table/Fig-2]; and the buccal and palatal surface between the two interdental contact points of 2^nd premolars [Table/Fig-3]. All measurements were carried out by the same operator and evaluated to the nearest of 0.01 mm [11].

[Table/Fig-1]:

The labial and palatal surface of central incisors.

[Table/Fig-2]:

The labial and palatal surface of canine.

[Table/Fig-3]:

The buccal and palatal surface of second premolar.

Statistical Package for the Social Sciences (SPSS) software (version 22.0; IBM, Armonk, NY) was used for statistical analysis. Measurements performed were six on three teeth on digital models and study models respectively. Mean and Standard Deviation (SD) were calculated. Unpaired t-test was used to check the significant difference in means of different parameters. Intra observer repeatability was evaluated using intraclass correlation analysis by repeating all measurements from five randomly selected models at a one month interval and all measurements were measured three time and mean average was recorded. Intra observer reliability of the measurements by a calibrated clinician resulted in a relatively small range of operator error with highly consistent repeated measurements. Intra observer error was assessed using Dalberg’s Formula. The level of significance was set at p-value <0.05 for all tests.

Mean biases, Standard Deviation (SD), t-value and p-values for the paired method comparisons are shown in [Table/Fig-4]. There were no significant differences between central incisors labial surface (p=0.845) and palatal surface (p=0.722) and the canine measurements labial surface (p=0.721) and palatal surface (p=0.544) on digital models compared to study models. Whereas significant difference was observed in premolar measurements on palatal surface (p=0.008) with higher accuracy seen in digital models to conventional models as the plaster models were considered as gold standard and the measurement of digital models were compared with them.

Studies have been reported in dental literature regarding the accuracy and validity of digital models compared to dental study models in respect to reproducibility of dental landmarks, measurements of mesio-distal width of teeth, arch width which are linear measurements and as diagnostic aid for treatment planning [8-10,15], using impressions scanned with LED or laser scanner or using intraoral scanners to generate digital files. Mack S et al,. study on 3D surface accuracy of intraoral scanners using curvilinear measurements was first to our knowledge to measure curve surface and brought to our notice importance of it which offers clinically more relevant information as opposed to linear measurements [11].

Impression scanning with LED scanner was used in this study because of its easy availability and feasibility for impression making to orthodontist. Intraoral scanners are expensive, time consuming procedure, and its findings might be affected by blood, saliva, and patient movements in-vivo [16,17], Also, differences in the light reflected back to the scanner can affect the scanner’s ability to determine the actual depth of an object accurately.

The STL generated from scanning the rubber based impressions where used with commercial software for curvilinear measurements, were the ruler was set to measure the true shortest path of a curve. This software allows moving the images around the three axis of rotation and magnifying the images. Magnification is an excellent advantage compared with the plaster models, because anatomic details can be viewed more accurately [11]. In some studies, it was demonstrated that it is possible to scan a patient intraorally to build satisfactory orthodontic clasps and other alloy components and embed them in a resin base without physical models [11,18]. One study concluded that the custom trays which were fabricated digitally are a closer match with the models of the patients as seen by the reserved impression space [19,11]. In comparison with handmade trays, digitally manufactured custom trays were more accurate.

Critical appraisal for this study is use of physical models which requires traditional manufacturing processes, impression materials would be necessary to obtain study model replications of the patient’s dentition and pouring with dental stone which are prone to deformation [20].

Hassan WNW et al,. compared measurements of digital calliper on plaster models vs 3D software on models scanned by structured-light scanner and concluded that measurements made on scanned models were in good agreement with manually made plaster models [21]. Fleming PS et al., concluded from its review article of orthodontic measurements on digital study models compared with plaster models that digital models offer a high degree of validity, similar result was found in this study [22]. The present study findings were similar to the findings of Mack S et al., study suggesting that STL files did not exhibit significant deformation or loss of information as evaluated by curvilinear measurements, and would provide a reliable surface area for appliance fabrication [11]. Significant difference was obtained in the premolars palatal region digital models obtained might contain small discrepancies this can be partially explained by registration errors of the patched 3D surfaces [11,23]. Gül Amuk N et al,. [24] compared digital models obtained from plaster dental model scanning and dental impression scanning and showed high accuracy and reliability. Intra observer reliability of the measurements by a calibrated clinician resulted in a relatively small range of operator error with highly consistent repeated measurements. Park SH et al,. measured the curved arch length and curved arch length discrepancy on model scanned digital model and intraoral scanned digital model which exhibited high validity when compared with the plaster models [25].

Clinical implications: As there is same accuracy between 3D curvilinear measurements on digital models generated from LED scanned impressions as compared with manual measurements on study models, so we can replace the plaster models with digital scanned impression models. This reduces the need for storage and maintenance problem and limits the physical damage to plaster models. The digital models can be stored in computer hard drive and potable CDs or on central server. There digital models are excellent tool for patient education that can be used to illustrate the improvement in treatment.

The sample size of the study was small. Only one type of scanning method was used. Further, with more sample examination more accurate and reliable results can be obtained. In future comparison of different scanning methods for its reliability to conventional models can be done.

The accuracy of measurements of curve surface on digital models was found similar to direct measurements on the dental casts with a caliper in relation to all measured tooth surface. Measurements of curve surface on digitised models showed higher accuracy in relation to premolar tooth as interdental embrasure areas can be accurately measured with digital measuring software where brass wire cannot reach manually. Clinically, digital models can be used for appliance fabrication and customisation of brackets for the respected patients. Different scanning options should also be checked for its reliability to produce curve surface.