Application of Multilevel Models to Morphometric Data. Part 1. Linear Models and Hypothesis Testing

Abstract

Morphometric data usually have a hierarchical structure (i.e., cells are nested within patients), which should be taken into consideration in the analysis. In the recent years, special methods of handling hierarchical data, called multilevel models (MM), as well as corresponding software have received considerable development. However, there has been no application of these methods to morphometric data yet. In this paper we report our first experience of analyzing karyometric data by means of MLwiN – a dedicated program for multilevel modeling. Our data were obtained from 34 follicular adenomas and 44 follicular carcinomas of the thyroid. We show examples of fitting and interpreting MM of different complexity, and draw a number of interesting conclusions about the differences in nuclear morphology between follicular thyroid adenomas and carcinomas. We also demonstrate substantial advantages of multilevel models over conventional, single‐level statistics, which have been adopted previously to analyze karyometric data. In addition, some theoretical issues related to MM as well as major statistical software for MM are briefly reviewed.

Protective effect of annexin‐A1 against irreversible damage to cavernous tissue after cavernous nerve injury in the rat

“…It is also known that changes in nuclear volume reflect the functional status of cells. When cells are in full physiological activity the nuclear volume is higher than when they are at rest [22]. This is probably the result of changes induced by the synthesis of DNA, RNA or proteins related to the cell maintenance process [21,22].…”

Classifying tissue samples from measurements on cells with within-class tissue sample heterogeneity

“…There is limited statistical understanding of the best approach for such a classification problem (Tsybrovskyy and Berghold, 2003a;Schulerud and others, 1998). There have been 3 general approaches (i) classify the micro-level observations and use an ad hoc approach to classify the macro-level (Assailly and Zhang, 1989;Grohs and Husain, 1994;Korbelik and others, 1993), (ii) extract macro-level features from the micro-level measurements (Thiran and Macq, 1996;Lugli and others, 2007;Bashashati and Brinkman, 2009), and (3) use a statistical model that accounts for the nested structure of the data (Cadez and others, 1999;Swartz and others, 2005).…”

“…Tsybrovskyy and Berghold (1999) showed that ignoring the macro-level as unit of analysis leads to an increase in bias and falsely lowered p-values. A multilevel or hierarchical method is the best approach to this problem, yet methods for classification have not been well developed-see also Tsybrovskyy and Berghold (2003a). If the problem is simplified by selecting only one kind of statistical unit, using the patient as the unit of analysis is the only way to avoid biased p-values.…”

Η Διερεύνηση Της Σημασίας Τεχνικών Τεχνητής Νοημοσύνης Στην Αξιολόγηση Της Μορφομετρικής Ανάλυσης Επιχρισμάτων Διαγνωστικής Κυτταρολογίας

“…Επιπρόσθετα, είναι γνωστό ότι η κυτταρολογική εξέταση είναι μια χρονοβόρα διαδικασία, η οποία βασίζεται κυρίως σε υποκειμενικά κριτήρια, αφού δεν χρησιμοποιούνται σε μεγάλο βαθμό μετρητικές διαδικασίες (και επομένως πιο αντικειμενικά κριτήρια) στην καθημερινή διαγνωστική πρακτική. Για το λόγο αυτό την τελευταία δεκαετία έχουν αναπτυχθεί διάφορες τεχνικές (ανάμεσα στις οποίες και στατιστικές μέθοδοι) που σκοπό έχουν τη χρήση ποσοτικοποιημένων μετρήσεων στη διάγνωση [3,4,5,6,7,8,9,10,11,12]. Έτσι η διάγνωση στη νόσο του καρκίνου γίνεται όσο το δυνατόν περισσότερο αντικειμενική.…”

“…Στη διάγνωση του καρκίνου, σε διάφορες μορφές του και σε διάφορα όργανα, έχουν χρησιμοποιηθεί αρκετές από τις προαναφερθείσες τεχνικές στη διεθνή βιβλιογραφία [3,4,5,6,7,8,9,10,11,16,36,37]. Οι τεχνικές αυτές αφορούν κυρίως πολυμεταβλητές μεθόδους και οι προσπάθειες επικεντρώνονται κυρίως στο διαχωρισμό των κυττάρων μεταξύ καλοήθειας και κακοήθειας (ή κάποια είδη αλλοιώσεων από κάποια άλλα σε ένα συγκεκριμένο όργανο).…”