Digital Theses Archive


Tesi etd-12202022-105841

Type of thesis
Novel technological solutions to engineer anatomic pathology processes
Scientific disciplinary sector
Istituto di Biorobotica - PHD IN BIOROBOTICA
relatore Prof. RICOTTI, LEONARDO
  • Anatomic pathology processes
  • indexing
  • laboratory automation
  • microtomy
  • virtual marker reconstruction
Exam session start date
Anatomic pathology is a branch of medicine that analyzes the structure of human tissues to reveal abnormalities, thus supporting clinicians in making a correct diagnosis and managing a patient&#39;s therapy. Pathologists identify and diagnose diseases by examining a piece of tissue through an optical microscope. Such a tissue can be extracted from a living body (through biopsies), or derived from a cadaver for post-mortem examinations (in autopsies). Once collected from the patient, the tissue undergoes several processes: tissue processing, embedding in paraffin, sectioning with a microtome, collection on a glass slide, staining to highlight the cell components and final diagnosis.<br>At present, over 90% of tests in an anatomic pathology laboratory (APL) are directed toward cancer diagnosis. The global cancer incidence is dramatically increasing: by 2040, the number of new cancer cases per year is expected to rise to 29.5 million. This forecast determines the need for a higher number of tissue-based diagnostic services. <br>The APL environment is much more complex than the one of other clinical/diagnostic laboratories, such as chemistry, hematology or bacteriology, in which the concept of total laboratory automation (TLA) has been already efficiently introduced. APLs are characterized by a broad spectrum of variables (e.g., sample type and container variability, presence of many different analytical processes, etc.). This considerable variability slows down the automation process. As a consequence, APL processes are mostly manually held today. Thus, they are inherently prone to errors, which can imply severe clinical consequences. <br>The main goal of my Ph.D. has been to propose novel technological solutions to automate anatomic pathology processes, also broadening the scientific knowledge behind such processes and building integrated platforms that could be used in the clinic, in the future.<br>The introduction of a new phase (the indexing process) has been necessary to standardize a process that usually is manually done by the technicians. This process has been inserted before the embedding in paraffin and arises from the need of an anatomopathologist to trace the coordinates of a point of interest for the tissue on the glass slide and transfer this information to the corresponding paraffin tissue block. Indeed, having a reference (fiducial marker) in the paraffin tissue block would allow carrying out further analyses, by picking the tissue sample and easily identifying the point of interest. An Indexing platform was developed in order to automate this newly introduced phase. It also required the development of another platform which was called Virtual Marker Reconstruction (VMR) and was introduced in two different moments, pre- and post-staining. During the staining the fiducial marker inserted with the indexing platform and present on pre-staining glass slide, is removed. Therefore, the VMR platform allows the virtual recontruction of the fiducial marker on the post-staining glass slide and saves it in a database, together with the image pre-staining.<br>Furthermore, the automation of the microtomy phase was analyzed in all its complexities. Microtomy is the phase in which thin sections of tissue (thickness: 2-10 µm) are cut from a paraffin tissue block, using a microtome. The first step of the sectioning consists of removing the initial paraffin layer that covers the tissue. This step is called Trimming. Once the tissue is correctly exposed, some sections are cut. This step is called Sectioning and aims at producing sections of good quality, namely without tears, wrinkles and folds. The two subphases of the microtomy process (Trimming and Sectioning) have been examined in depth and automatized by developing two different platforms. It was demonstrated that the two processes can be parallelized, and this can accelerate the process throughput.<br>All the platforms developed have an artificial intelligence on board able to simulate the technician’s experience and allowing a proper automation. The solutions proposed to automate the indexing process and the microtomy have been approved by the end-users (technicians and pathologists) and have been judged useful and innovative. <br>The efforts of this PhD thesis led to the filing of 2 patents, to the publication of 1 journal paper. Further 3 journal papers are also in preparation.