Skip to main content
  • Main menu
  • User menu
  • Search
  • English ▼
    • English
    • Afrikaans
    • Albanian
    • Amharic
    • Arabic
    • Armenian
    • Azerbaijani
    • Basque
    • Belarusian
    • Bengali
    • Bosnian
    • Bulgarian
    • Catalan
    • Cebuano
    • Chichewa
    • Chinese (Simplified)
    • Chinese (Traditional)
    • Corsican
    • Croatian
    • Czech
    • Danish
    • Dutch
    • Esperanto
    • Estonian
    • Filipino
    • Finnish
    • French
    • Frisian
    • Galician
    • Georgian
    • German
    • Greek
    • Gujarati
    • Haitian Creole
    • Hausa
    • Hawaiian
    • Hebrew
    • Hindi
    • Hmong
    • Hungarian
    • Icelandic
    • Igbo
    • Indonesian
    • Irish
    • Italian
    • Japanese
    • Javanese
    • Kannada
    • Kazakh
    • Khmer
    • Korean
    • Kurdish (Kurmanji)
    • Kyrgyz
    • Lao
    • Latin
    • Latvian
    • Lithuanian
    • Luxembourgish
    • Macedonian
    • Malagasy
    • Malay
    • Malayalam
    • Maltese
    • Maori
    • Marathi
    • Mongolian
    • Myanmar (Burmese)
    • Nepali
    • Norwegian
    • Pashto
    • Persian
    • Polish
    • Portuguese
    • Punjabi
    • Romanian
    • Russian
    • Samoan
    • Scottish Gaelic
    • Serbian
    • Sesotho
    • Shona
    • Sindhi
    • Sinhala
    • Slovak
    • Slovenian
    • Somali
    • Spanish
    • Sudanese
    • Swahili
    • Swedish
    • Tajik
    • Tamil
    • Telugu
    • Thai
    • Turkish
    • Ukrainian
    • Urdu
    • Uzbek
    • Vietnamese
    • Welsh
    • Xhosa
    • Yiddish
    • Yoruba
    • Zulu

Main menu

  • Home
  • Content
    • Current
      • JNMT Supplement
    • Ahead of print
    • Past Issues
    • Continuing Education
    • JNMT Podcast
    • SNMMI Annual Meeting Abstracts
  • Subscriptions
    • Subscribers
    • Rates
    • Journal Claims
    • Institutional and Non-member
  • Authors
    • Submit to JNMT
    • Information for Authors
    • Assignment of Copyright
    • AQARA Requirements
  • Info
    • Reviewers
    • Permissions
    • Advertisers
    • Corporate & Special Sales
  • About
    • About Us
    • Editorial Board
    • Contact Information
  • More
    • Alerts
    • Feedback
    • Help
    • SNMMI Journals
  • SNMMI
    • JNMT
    • JNM
    • SNMMI Journals
    • SNMMI

User menu

  • Subscribe
  • My alerts
  • Log in
  • Log out
  • My Cart

Search

  • Advanced search
Journal of Nuclear Medicine Technology
  • SNMMI
    • JNMT
    • JNM
    • SNMMI Journals
    • SNMMI
  • Subscribe
  • My alerts
  • Log in
  • Log out
  • My Cart
Journal of Nuclear Medicine Technology

Advanced Search

English ▼
  • English
  • Afrikaans
  • Albanian
  • Amharic
  • Arabic
  • Armenian
  • Azerbaijani
  • Basque
  • Belarusian
  • Bengali
  • Bosnian
  • Bulgarian
  • Catalan
  • Cebuano
  • Chichewa
  • Chinese (Simplified)
  • Chinese (Traditional)
  • Corsican
  • Croatian
  • Czech
  • Danish
  • Dutch
  • Esperanto
  • Estonian
  • Filipino
  • Finnish
  • French
  • Frisian
  • Galician
  • Georgian
  • German
  • Greek
  • Gujarati
  • Haitian Creole
  • Hausa
  • Hawaiian
  • Hebrew
  • Hindi
  • Hmong
  • Hungarian
  • Icelandic
  • Igbo
  • Indonesian
  • Irish
  • Italian
  • Japanese
  • Javanese
  • Kannada
  • Kazakh
  • Khmer
  • Korean
  • Kurdish (Kurmanji)
  • Kyrgyz
  • Lao
  • Latin
  • Latvian
  • Lithuanian
  • Luxembourgish
  • Macedonian
  • Malagasy
  • Malay
  • Malayalam
  • Maltese
  • Maori
  • Marathi
  • Mongolian
  • Myanmar (Burmese)
  • Nepali
  • Norwegian
  • Pashto
  • Persian
  • Polish
  • Portuguese
  • Punjabi
  • Romanian
  • Russian
  • Samoan
  • Scottish Gaelic
  • Serbian
  • Sesotho
  • Shona
  • Sindhi
  • Sinhala
  • Slovak
  • Slovenian
  • Somali
  • Spanish
  • Sudanese
  • Swahili
  • Swedish
  • Tajik
  • Tamil
  • Telugu
  • Thai
  • Turkish
  • Ukrainian
  • Urdu
  • Uzbek
  • Vietnamese
  • Welsh
  • Xhosa
  • Yiddish
  • Yoruba
  • Zulu
  • Home
  • Content
    • Current
    • Ahead of print
    • Past Issues
    • Continuing Education
    • JNMT Podcast
    • SNMMI Annual Meeting Abstracts
  • Subscriptions
    • Subscribers
    • Rates
    • Journal Claims
    • Institutional and Non-member
  • Authors
    • Submit to JNMT
    • Information for Authors
    • Assignment of Copyright
    • AQARA Requirements
  • Info
    • Reviewers
    • Permissions
    • Advertisers
    • Corporate & Special Sales
  • About
    • About Us
    • Editorial Board
    • Contact Information
  • More
    • Alerts
    • Feedback
    • Help
    • SNMMI Journals
  • Watch or Listen to JNMT Podcast
  • Visit SNMMI on Facebook
  • Join SNMMI on LinkedIn
  • Follow SNMMI on Twitter
  • Subscribe to JNMT RSS feeds
Research ArticleAI/Advanced Image Analysis

Remodeling 99mTc-Pertechnetate Thyroid Uptake: Statistical, Machine Learning, and Deep Learning Approaches

Geoffrey M. Currie and Basit Iqbal
Journal of Nuclear Medicine Technology June 2022, 50 (2) 143-152; DOI: https://doi.org/10.2967/jnmt.121.263081
Geoffrey M. Currie
1Charles Sturt University, Wagga Wagga, Australia, and Baylor College of Medicine, Houston, Texas; and
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
Basit Iqbal
2Gujranwala Institute of Nuclear Medicine and Radiotherapy, Gujranwala, Pakistan
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • Article
  • Figures & Data
  • Info & Metrics
  • PDF
Loading

Article Figures & Data

Figures

  • Tables
  • Figure
    • Download figure
    • Open in new tab
    • Download powerpoint
  • FIGURE 1.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 1.

    Intuitive, but sometimes inaccurate, visual evaluation of thyroid status relative to salivary gland activity. (Left) Salivary gland activity exceeding thyroid gland activity suggests hypothyroidism. (Middle) Salivary gland activity and thyroid gland activity being similar (within same scale) suggests euthyroidism. (Right) Salivary gland activity not being apparent relative to thyroid activity suggests hyperthyroidism. All images were obtained with 99mTc-pertechnetate using high-resolution, parallel-hole imaging.

  • FIGURE 2.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 2.

    CNN architecture. 2D = 2-dimensional; ReLU = rectified linear unit.

  • FIGURE 3.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 3.

    Three example patients (top, middle, and bottom) with black on white (left), white on black (center), and magnitude spectrum from Fourier transformation (right) used as inputs for CNN.

  • FIGURE 4.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 4.

    (Left) Ternary biochemical status classification against thyroid uptake. (Right) Broader biochemical status classification against thyroid uptake. Horizontal line represents overall mean, and diamonds represent class mean and 95% CIs.

  • FIGURE 5.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 5.

    Final architecture of trained and validated neural network. TSH = thyroid-stimulating hormone.

  • FIGURE 6.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 6.

    Cumulative gain chart demonstrating maximum separation of positive and negative curves to provide cumulative gain score of 0.8 and instances ratio of 0.4 (arrow).

  • FIGURE 7.
    • Download figure
    • Open in new tab
    • Download powerpoint
    FIGURE 7.

    Various scintigraphic appearances of thyroid pathology using parallel-hole (high-resolution) collimation and 99mTc-pertechnetate. MNG = multinodular goiter.

Tables

  • Figures
    • View popup
    TABLE 1.

    Biochemical Stratification of Patient Studies and Findings (1–6,12,13)

    Free T3 (2–7 pmol/L*)Free T4 (12–30 pmol/L*)Thyroid-stimulating hormone (0.45–4.5 μIU/mL*)Biochemical status99mTc uptake (%)Comment on uptake reference range
    HighHighLowHyperthyroidism>4.50% FN rate
    NormalNormalLowSubclinical hyperthyroidism<4.5 including <0.45 or absent0% TP, comprised FN or FP hypothyroidism
    HighNormalLowT3 toxicosis>4.5 or <0.45FP hypothyroidism
    NormalHighLowThyroiditisNo cases
    LowLowLowSecondary hypothyroidismNo cases
    NormalNormalHighSubclinical hypothyroidism>0.45 but <4.5100% FN
    Low or normalLowHighPrimary hypothyroidism>0.45 and in over 50% of cases >4.5100% FN
    NormalNormalNormalEuthyroid<4.5%9% FP rate (6% hyperthyroid, 3% hypothyroid)
    • * Reference range.

    • View popup
    TABLE 2.

    CNN Architecture, Activations, and Parameters

    LayerNameActivationsParameters
    1Tensor input layer[725,725,3]
    22D convolution layer[239,239,64]Weights [11,11,3,64], bias [1,1,64]
    3Batch normalization[239,239,64]Offset and scale [1,1,64]
    4ReLU layer[239,239,64]
    5Max pooling layer[119,119,64]Size [3,3], stride [2,2], padding [0,0,0,0]
    62D convolution layer[40,40,128]Weights 5,5,64,128], bias [1,1,128]
    7Batch normalization[40,40,128]Offset and scale [1,1,128]
    8ReLU layer[40,40,128]
    9Max pooling layer[19,19,128]Size [3,3], stride [2,2], padding [0,0,0,0]
    102D convolution layer[19,19,256]Weights [3,3,128,256], bias [1,1,256]
    11Batch normalization[19,19,256]Offset and scale [1,1,256]
    12ReLU layer[19,19,256]
    13Max pooling layer[9,9,256]Size [3,3], stride [2,2], padding [0,0,0,0]
    142D convolution layer[9,9,192]Weights [3,3,256,192], Bias [1,1,192]
    15Batch normalization[9,9,192]Offset and scale [1,1,192]
    16ReLU layer[9,9,192]
    17Max pooling layer[4,4,192]Size [3,3], stride [2,2], padding [0,0,0,0]
    182D convolution layer[4,4,192]Weights [3,3,256,192], bias [1,1,192]
    19Batch normalization[4,4,192]Offset and scale [1,1,192]
    20ReLU layer[4,4,192]
    21Max pooling layer[1,1,192]Size [3,3], stride [2,2], padding [0,0,0,0]
    22Fully connected layer[1,1,192]Weights [192,192], bias [192,1]
    23ReLU layer[1,1,192]
    24Dropout layer[1,1,192]0.5
    25Fully connected layer[1,1,86]Weights [86,192], bias [86,1]
    26ReLU layer[1,1,86]
    27Dropout layer[1,1,86]0.5
    28Fully connected layer[1,1,2]Weights [2,86], bias [2,1]
    29Softmax layer[1,1,2]
    30Classification layerCross entropy loss function
    • 2D = 2-dimensional; ReLU = rectified linear unit.

    • View popup
    TABLE 3.

    Ternary Classification of Thyroid Function Based on Various Published Reference Ranges

    Reference rangeEuthyroidHyperthyroidHypothyroidReference
    0.45%–4.5%67.5%26.8%7.7%6
    0.4%–1.7%35.0%61.0%4.0%3
    0.4%–4.0%65.0%31.0%4.0%4
    0.3%–3.4%57.7%38.2%4.1%2
    0.2%–2.0%43.1%52.8%4.1%5
    Biochemical status53.1%27.1%19.8%*11
    Salivary classification44.8%50.0%5.2%—
    Physician visual rating51.0%43.8%5.2%—
    Physician rating with uptake value64.6%29.2%6.3%—
    • *15.6% were hypothyroid without suppression of uptake (2.1% autonomous, 2.1% secondary hypothyroidism, 11.5% primary hypothyroidism, and 4.2% subclinical hypothyroidism).

    • View popup
    TABLE 4.

    Key Variables

    VariableMean95% CI
    Total count ratio of right-lobe activity to left-lobe activity1.51.03–2.02
    CPP ratio of right-lobe activity to left-lobe activity1.290.98–1.60
    Area33.8 cm231.1–36.5
    Size, right3092 pixels2,848–3,340
    Size, left2937 pixels2,662–3,212
    Ratio of thyroid to background4.063.43–4.69
     Right4.01 CPP3.49–4.52
     Left4.08 CPP3.28–4.89
    Ratio of dose to total counts4.853.44–6.26
    FT421.1 pmol/L18.1–24.2
    FT37.1 pmol/L5.1–9.1
    Thyroid-stimulating hormone4.2 pmol/L2.3–6.1
    • View popup
    TABLE 5.

    Ternary Classification of Thyroid Function Based on Recall Against Biochemical Status

    Reference rangeEuthyroidHyperthyroid*HypothyroidAccuracy†
    0.45%–4.5%71.4%66.6% (100%)0%82.6%
    0.4%–1.7%49.0%74.1% (94.1%)0%51.0%
    0.4%–4.0%86.3%63.0% (94.1%)0%77.1%
    0.3%–3.4%74.5%63.0% (94.1%)0%68.8%
    0.2%–2.0%58.8%74.1% (94.1%)0%59.4%
    Salivary classification62.7%70.3% (94.1%)0%61.4%
    Physician rating72.5%63.0% (89.5%)0%61.0%
    Physician rating with uptake88.2%70.3% (100%)0%82.3%
    • * Data in parentheses exclude subclinical hyperthyroidism and T3 toxicosis.

    • ↵†Binary accuracy for reference to Table 6.

    • Accuracy is also provided for binary classification.

    • View popup
    TABLE 6.

    Triplicate Training and Validation Binary Results (Hyperthyroid or Not Hyperthyroid) for 30-Layer CNN Architecture

    Input tensorTraining accuracyTraining lossValidation accuracyValidation lossMean validation accuracyBinary accuracy
    White on black82.1%0.42075.9%0.53680.5%
    94.0%0.22579.3%0.602
    91.0%0.21886.2%0.414
    Black on white83.6%0.38382.8%0.40578.2%
    80.6%0.45272.4%0.544
    91.0%0.23279.3%0.690
    Magnitude spectrum76.1%0.45975.9%0.53075.9%
    74.6%0.50872.4%0.542
    85.1%0.30679.3%0.380
    Mean84.2%0.35678.2%0.516
    Initial 25-layer CNN69.0%
    Conventional metrics
     Normal cutoff, 4.5%82.6%
     Normal cutoff, 4.0%77.1%
     Salivary classification61.5%
     Physician rating61.0%
     Physician rating with uptake82.3%
    • Corresponding binary accuracies of best-performing thyroid uptake cutoffs, visual classification against salivary activity relative to thyroid activity, and physician rating are included for comparison.

PreviousNext
Back to top

In this issue

Journal of Nuclear Medicine Technology: 50 (2)
Journal of Nuclear Medicine Technology
Vol. 50, Issue 2
June 1, 2022
  • Table of Contents
  • About the Cover
  • Index by author
  • Complete Issue (PDF)
Print
Download PDF
Article Alerts
Email Article
Citation Tools
Share
Remodeling 99mTc-Pertechnetate Thyroid Uptake: Statistical, Machine Learning, and Deep Learning Approaches
Geoffrey M. Currie, Basit Iqbal
Journal of Nuclear Medicine Technology Jun 2022, 50 (2) 143-152; DOI: 10.2967/jnmt.121.263081
Twitter logo Facebook logo LinkedIn logo Mendeley logo
  • Tweet Widget
Bookmark this article

Jump to section

  • Article
    • Visual Abstract
    • Abstract
    • MATERIALS AND METHODS
    • RESULTS
    • DISCUSSION
    • CONCLUSION
    • DISCLOSURE
    • ACKNOWLEDGMENT
    • Footnotes
    • REFERENCES
  • Figures & Data
  • Info & Metrics
  • PDF

Related Articles

  • No related articles found.
  • PubMed
  • Google Scholar

Cited By...

  • Its Time to Gather and Share
  • Google Scholar

Similar Articles

  • PET/CT Case Series: Unmasking the Mystery of Cardiac Sarcoidosis
  • Cardiac Sarcoidosis Imaging: PET
  • Patient Motion During Cardiac PET Imaging
  • PET Study Guide
  • PET/MRI Assessment of Acute Cardiac Inflammation 1 Month After Left-Sided Breast Cancer Radiation Therapy
See more

Keywords

  • thyroid uptake
  • hyperthyroidism
  • machine learning
  • deep learning
  • artificial intelligence
SNMMI

© 2025 SNMMI

Powered by HighWire
Alerts for this Article
Sign In to Email Alerts with your Email Address
Email this Article

Thank you for your interest in spreading the word on Journal of Nuclear Medicine Technology.

NOTE: We only request your email address so that the person you are recommending the page to knows that you wanted them to see it, and that it is not junk mail. We do not capture any email address.

Enter multiple addresses on separate lines or separate them with commas.
Remodeling 99mTc-Pertechnetate Thyroid Uptake: Statistical, Machine Learning, and Deep Learning Approaches
(Your Name) has sent you a message from Journal of Nuclear Medicine Technology
(Your Name) thought you would like to see the Journal of Nuclear Medicine Technology web site.
Citation Tools
Remodeling 99mTc-Pertechnetate Thyroid Uptake: Statistical, Machine Learning, and Deep Learning Approaches
Geoffrey M. Currie, Basit Iqbal
Journal of Nuclear Medicine Technology Jun 2022, 50 (2) 143-152; DOI: 10.2967/jnmt.121.263081

Citation Manager Formats

  • BibTeX
  • Bookends
  • EasyBib
  • EndNote (tagged)
  • EndNote 8 (xml)
  • Medlars
  • Mendeley
  • Papers
  • RefWorks Tagged
  • Ref Manager
  • RIS
  • Zotero

We use cookies on this site to enhance your user experience

By clicking any link on this page you are giving your consent for us to set cookies.