Multi-Dimensional Morphological Characterization and Drug Effects of Tumor Organoids Based on OCT and Machine Learning

Chuanwei Mao; Shanshan Yang; Xiao Liang; Ling Wang; Ming en Xu

doi:10.3788/CJL240595

Chinese Journal of Lasers, Volume. 51, Issue 15, 1507108(2024)

Multi-Dimensional Morphological Characterization and Drug Effects of Tumor Organoids Based on OCT and Machine Learning

Chuanwei Mao¹, Shanshan Yang^1,2、*, Xiao Liang³, Ling Wang^1,2, and Ming en Xu^1,2

Author Affiliations

¹School of Automation, Hangzhou Dianzi University, Hangzhou 310018, Zhejiang , China

²Key Laboratory of Medical Information and 3D Bioprinting of Zhejiang Province, Hangzhou 310018, Zhejiang , China

³General Surgery Department, Sir Run Run Shaw Hospital Affiliated to Zhejiang University School of Medicine, Hangzhou 310018, Zhejiang , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Optical coherence tomography imaging device. (a) A physical image of the OCT prototype, with 1, 2, and 3 representing the interference module, the three-axis mobile module carrying the imaging probe, and the well plate placement area, respectively; (b) a physical image of the whole organoid imaging analyzer

Download full size

Fig. 2. Organoid imaging results. (a1) 3D OCT imaging; (a2) 2D OCT imaging; (b1) segmentation result of 3D OCT image; (b2) segmentation result of 2D OCT image; (c1)‒(c9) nine single organoids in the organoid cluster

Download full size

Fig. 3. Frequency distributions of morphological parameters (solid volume, surface area, and cavity volume) for single organoids before and after undersampling

Download full size

Fig. 4. Clustering analysis results and bright-field microscopy images of organoid clusters in different wells before and after dosing. (a1)‒(d1) Clustering analysis results of organoid clusters in each well before dosing, where the blue channel represents category 1, the green channel represents category 2, and the orange channel represents category 3; (a2)‒(d2) bright-field microscopy images of organoid clusters in each well before dosing; (a3)‒(d3) clustering analysis results of organoid clusters in each well after dosing; (a4)‒(d4) bright-field microscopy images of organoid clusters in each well after dosing

Download full size

Table 0. [in Chinese]

View table

Table 0. [in Chinese]

Serial number	Concentration of icaritin /（μmol/L）	Well number	Composite score of the growth level F		Relative growth value	ATP value on Day5
Serial number	Concentration of icaritin /（μmol/L）	Well number	F_Day3	F_Day5	F_Day5-F_Day3	ATP value on Day5
19	Control	E11	0.057	0.838	0.781	14700000
20		F11	-0.536	0.292	0.828	11180000
21		F2	0.574	2.076	1.502	18240000
22		F6	-0.488	0.872	1.360	20320000
23		F8	-0.716	0.998	1.714	17170000
24		F9	-0.282	0.338	0.620	15830000

Table 1. Organoid groups and drug regimens

View table

Table 1. Organoid groups and drug regimens

Group number	Well number	Drug	Concentration
Group1_1	B2, B3, B4, C2, C3, C4	Icaritin	20 μmol/L
Group1_2	B5, B6, B7, C5, C6, C7		40 μmol/L
Group1_3	B8, B9, B10, C8, C9, C10		80 μmol/L
Group2_1	D2, D3, D4, E2, E3, E4	Cisplatin	20 μg/mL
Group2_2	D5, D6, D7, E5, E6, E7		50 μg/mL
Group2_3	D8, D9, D10, E8, E9, E10		100 μg/mL
Group3	F2, F3, F4, F5, F6, F7, F8 F9, F10, F11, B11, C11, D11, E11	Control

Table 2. Morphological parameters of the nine organoids in Fig. 2(c)

View table

Table 2. Morphological parameters of the nine organoids in Fig. 2(c)

Organoid number	Solid volume / $μ m^{3}$	Surface area / $μ m^{2}$	Cavity volume / $μ m^{3}$	Cyst thickness / $μ m$	Sphericity /arb.units	Longest-axis length / $μ m$	Shortest-axis length / $μ m$
c1	9209375	541925	8929875	17	0.62	797	340
c2	7513875	458225	7867250	16	0.65	528	305
c3	3889375	277800	3297875	14	0.65	468	250
c4	3811625	244800	1374125	16	0.59	554	205
c5	3162125	241875	0	13	0.43	701	265
c6	2976250	221450	991125	13	0.55	459	265
c7	2588250	212700	511500	12	0.48	554	180
c8	2008250	188875	321000	11	0.45	390	210
c9	2539875	185775	153875	14	0.50	546	145

Table 3. Statistical results of morphological parameters based on clustering (B9)

View table

Table 3. Statistical results of morphological parameters based on clustering (B9)

Dosing or not	Category No.	Number	Solid volume average / $μ m^{3}$	Surface area average / $μ m^{2}$	Cavity volume / $μ m^{3}$
Before dosing	1	284	13378±7342	2442±1202	0
	2	151	209267±282970	22324±22596	0
	3	9	4188778±2334620	285936±119205	2605181±2243600
After dosing	1	226	13627±7947	2530±1335	0
	2	172	253583±376374	26592±30998	0
	3	8	4972031±2989793	274566±105665	0
Dosing or not	Category No.	Longest-axis length average / $μ m$	Shortest-axis length average / $μ m$	Cyst thickness average / $μ m$	Sphericity average /arb.units
Before dosing	1	61±17	16±7	0	1.15±0.15
	2	142±70	60±70	0	0.76±0.12
	3	555±118	241±58	14±2	0.55±0.08
After dosing	1	63±21	14±6	0	1.13±0.15
	2	146±82	61±38	0	0.74±0.13
	3	509±97	230±57	17±4	0.50±0.07

Table 4. Principal component load matrix

View table

Table 4. Principal component load matrix

Selected features (eigenvalue >1)	PCA1	PCA2	PCA3	PCA4
Number 1 (X₁)	-0.375	-0.586	0.470	-0.308
Solid volume 1 (X₂)	-0.356	-0.218	0.038	0.796
Axial ratio 1 (X₃)	0.346	-0.264	0.421	0.124
Number 2 (X₄)	-0.062	-0.742	0.386	-0.418
Solid volume 2 (X₅)	-0.608	0.584	0.505	-0.087
Surface area 2 (X₆)	-0.592	0.596	0.530	-0.022
Sphericity 2 (X₇)	0.654	-0.651	-0.290	0.046
Number 3 (X₈)	0.440	0.257	0.156	-0.330
Solid volume 3 (X₉)	0.767	0.224	0.420	0.283
Surface area 3 (X₁₀)	0.850	0.231	0.338	0.118
Cavity volume 3 (X₁₁)	0.604	0.577	-0.324	-0.131
Cystic thickness 3 (X₁₂)	-0.699	-0.222	-0.119	0.280
Sphericity 3 (X₁₃)	-0.435	0.236	-0.720	-0.245

Table 5. Twenty-four sets of verification data for the composite scores of PDOs growth level

View table

Table 5. Twenty-four sets of verification data for the composite scores of PDOs growth level

Serial number	Concentration of icaritin /（μmol/L）	Well number	Composite score of the growth level F		Relative growth value	ATP value on Day5
Serial number	Concentration of icaritin /（μmol/L）	Well number	F_Day3	F_Day5	F_Day5-F_Day3	ATP value on Day5
1	80	B8	0.419	0.019	-0.400	312700
2		B9	-0.278	-0.690	-0.412	140300
3		B10	-0.106	0.053	0.159	601300
4		C10	0.237	-0.507	-0.744	211800
5		C8	0.212	-0.439	-0.651	211300
6		C9	0.268	-0.203	-0.471	465900
7	40	B5	-0.315	-0.567	-0.252	511900
8		B6	-0.160	-0.007	0.153	404500
9		B7	0.074	-0.239	-0.313	403900
10		C5	-0.176	-0.104	0.072	330900
11		C6	0.102	-0.080	-0.182	238900
12		C7	0.405	-0.340	-0.745	682100
13	20	B2	-0.134	0.043	0.177	5391000
14		B3	-0.192	-0.118	0.074	6538000
15		B4	-0.025	0.129	0.154	7103000
16		C2	0.318	0.420	0.102	6336000
17		C3	0.218	0.322	0.105	8336000
18		C4	-0.362	-0.420	-0.058	6800000

Tools

Get Citation

Copy Citation Text

Chuanwei Mao, Shanshan Yang, Xiao Liang, Ling Wang, Ming en Xu. Multi-Dimensional Morphological Characterization and Drug Effects of Tumor Organoids Based on OCT and Machine Learning[J]. Chinese Journal of Lasers, 2024, 51(15): 1507108

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Biomedical Optical Imaging

Received: Feb. 19, 2024

Accepted: May. 7, 2024

Published Online: Jun. 28, 2024

The Author Email: Shanshan Yang (yangshan@hdu.edu.cn)

DOI:10.3788/CJL240595

CSTR:32183.14.CJL240595

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology