**4. Discussion and conclusion**

The presented measurement system and approach contain three main sources of potential error: (1) the accuracy of the 3D stereo‐photogrammetric capture, (2) operator‐dependent skin marker positioning and (3) data processing.

The first instance is no more an issue because present‐day high‐resolution stereo‐photogram‐ metric systems provide very high accuracy in 3D reconstruction of marker position (0.3–0.4‐mm range on a 3‐m wide by 3‐m deep by 2‐m high working volume in entry‐level standard 6TVC 0.3 Mpix GOALS system) using well‐established and easy‐to‐manage mature stereo‐photogram‐ metric calibration procedure.

The second source of potential error was the cause for some initial scepticism that this meth‐ odology had to suffer from. In fact, notwithstanding characterized by a series of advantages, doubts about such an approach were mostly due to the improper belief of a certain inaccuracy of obtained results because of manual cutaneous positioning of the passive retro‐reflective markers on the palpable bony landmarks. Additional concern was linked to supposed time‐ consuming nature of the marker‐positioning process. However, studies [24, 43–45] showed that, obviously, operators need to be trained, as for each imaging‐based technique but, when they reach the necessary skill, it has been demonstrated, via X‐ray and MRI, that they are able to perform by palpation accurate marker positioning on pre‐defined bony prominences. In particular on the spine, intra‐ and inter‐examiner errors resulted relatively low with respect to subject postural adjustment variability. Furthermore, the same studies confirmed that, when the operators are trained, the marker placement takes only few minutes, in addition being a substantial first stage of examination that helps as guidance for the rest of the analysis.

Indeed, the third source of error, that is, data processing, could be a major source of data corruption if improper signal‐processing techniques are used. Prior work by D'Amico et al. [7] originally introduced an efficient and specially devised numerical processing technique, showing maximal error of less than 1° (approximately) for the Cobb‐computed angle on a curve of about 65°. A subsequent study on scoliotic patients [17] presented an in vivo com‐ parison between frontal plane Cobb angles computed using stereo‐photogrammetry and AP (anterior‐posterior) X‐ray images. Results showed that Cobb angles measured via X‐ray matched those accessed via a stereo‐photogrammetric system.

(of spine and pelvis sagittal angles) and flexion torques, the healthy subject presents with respect to the pathological one, both a noticeably lower trunk torques (especially at S1–L5 low lumbar levels) and a wider range of spinal angles, with a different pattern showing different

**Figure 15.** Forward bending results for healthy subject and LBP patient. 3D skeleton reconstruction at maximum forward trunk flexion and measured SEMG activity during the forward‐bending test (upper‐row panels). Comparison healthy subject versus LBP patient on spine and pelvis measured sagittal ROMs (lower‐left panels) and flexion torques patterns

In particular, the LBP patient shows a lower mobility in the lumbar spine with a decrease of the ROM quantified in more than 25° lordosis angle value at maximum forward flexion, compen‐ sated by a higher mobility in the pelvis that presents an increase of the ROM quantified in more than 10°at maximum forward flexion. Conversely, thoracic spine mobility ranges are compara‐ ble in the two subjects. Such a condition reveals a stiffer behaviour in the lumbar spine of the LBP subject that could be connected to the still persistent FRP stop. In fact, SEMG recordings show, in the healthy subject, a clear presence of FRP as expected, while LBP subject still shows a full contraction activity in all the considered paravertebral muscles, that is, FRP stop. Further studies are currently in progress to determine, on a larger LBP population, if such described FRP stop could relate to modified stiffer pattern and reduced lumbar mobility as observed in this case.

The presented measurement system and approach contain three main sources of potential error: (1) the accuracy of the 3D stereo‐photogrammetric capture, (2) operator‐dependent skin

strategies in task execution.

(lower‐right panels), respectively.

38 Innovations in Spinal Deformities and Postural Disorders

**4. Discussion and conclusion**

marker positioning and (3) data processing.

The advantages of the presented measurement system and approach are related to the ability to quantitatively capture both static and dynamic body postural expressions, including spinal shape, using 3D imaging of skeleton posture. It can achieve this for a static‐erect posture and related oscillations, as well as additionally for the eventual behaviour of body segments dur‐ ing movement. This methodology has no harmful effects, so it provides a 'natural' approach for both capturing and monitoring the progression of pathology and/or the treatment out‐ comes. Moreover, researchers and clinicians can analyse the upright standing posture in real, neutral and unconstrained conditions, by obtaining, in a rapid and automatic way, a large number of useful 3D and 2D anatomical/biomechanical/clinical parameters. In addition, all possible postural characteristics are measured simultaneously and at a very fast frame rate. These features allow investigators to perform multiple different postural tests, within a same session, and to directly compare them in a statistically reliable way being the final assess‐ ment based always on averaged data either in static or in dynamic analysis. In addition, it allows a multi‐factorial‐multi‐sensors approach that provides the capability to fully integrate all the measurement data into a unified view to correlate morphological‐kinematic character‐ istics to full‐functional evaluation and to use such results for clinical aims. When the analysis is focused only on 3D posture complemented by bending tests, multi‐factorial clinical and biomechanical results (including the analysis and comparisons of standing‐erect posture taken in different conditions) are immediately available, in the form of comprehensive report, both to the operator and to the subject at the end of acquisition session. In general, the mean duration of such analysis using the GOALS system lasts from 30 to 45 min. A longer duration is necessary when other more complex motor tasks (gait, run, FRP, etc.) are included in the acquisition session. The case studies presented in this chapter are only exemplary summariz‐ ing examples of a wider research activity that our group has carried on in more than 25 years. The obtained results demonstrated such methodology as being entirely suitable and effec‐ tive for clinical application. This approach was not developed or intended to replace radio‐ graphs, from which much more information than spine morphology can be drawn. However, depending on the specific clinical purposes, this methodology could be used during screening and follow‐up to reduce patient irradiation, evaluation time and cost.

This framework is extremely useful as a baseline reference at the diagnostic stage not only lim‐ ited to orthopaedic field but its functional‐biomechanical analysis capability can be naturally extended in neurological disorders. For this reason, it could be extremely useful at the stage of designing and developing treatment programmes in rehabilitation, in planning of ortho‐ paedic surgical procedures and in monitoring the progression of pathology and/or treatment outcomes in any kind of postural disorders and pathologies. Additionally, it can be used ben‐ eficially for the study of postural characteristics connected to different sports activities as well as to evaluate postural change during growth and/or ageing processes. The GOALS system is under continuous development and improvement. Thousands of patients have been analysed and followed up with this methodology, identifying and precisely differentiating pathologi‐ cal patterns proving to be a type of analysis well accepted both by clinicians and by patients.
