TCP and NTCP Comprehensive Study Notes

Radiotherapy Treatment Aims & Professional Roles

  • Primary goal of radiotherapy (R/T)- Deliver a sufficiently high dose to eradicate tumour tissue.

    • Spare surrounding normal tissue as much as possible.

  • Key professionals- Radiation Oncologist (RO)

    • Prescribes course of treatment balancing tumour control & toxicity.

    • Radiation Therapist (RT)

    • Creates one or more computer-optimised treatment plans.

    • Computes dosimetry options for RO approval.

    • Medical Physicist (MP)

    • Verifies physical & biological accuracy of plans.

Historical Plan Evaluation & Dose-Volume Histograms (DVHs)

  • Early plan assessment relied on RO clinical experience only- Lacked practicality, consistency & objectivity.

  • Introduction of computerised DVH tools allowed quantitative plan comparison.- DVH = histogram of dose versus percentage (or absolute) volume of a contoured structure.

    • RO selects plan that maximises tumour dose while keeping each organ-at-risk (OAR) below tolerance thresholds.

  • Reference: instructional YouTube video (link supplied in transcript).

Limitations of DVHs

  • No spatial information – cannot identify where hot/cold spots reside within anatomy.

  • Ignore fractionation & radiobiology – radiosensitivity and repair kinetics not represented.

  • Contour-volume dependence - Different margins (e.g. 1 cm inside beam edge vs 2 cm outside) alter the DVH despite same patient & plan.

  • Result: Need additional biological metrics beyond purely geometric DVH.

Radiotherapy Volume Definitions (ICRU)

  • GTV (Gross Tumour Volume) – palpable/visible tumour (contains some normal tissue elements).

  • CTV (Clinical Target Volume) – GTV + region of potential microscopic spread; intermixed normal & malignant cells.

  • ITV (Internal Target Volume) – CTV plus internal motion margin.

  • PTV (Planning Target Volume) – ITV plus setup margins to account for geometric uncertainties.

  • Treated Volume / Irradiated Volume – volumetric region actually receiving significant dose; consists exclusively of normal tissue outside PTV.

  • Reality: Some normal tissue irradiation is inevitable → dose-limiting side-effects.

  • Modern advances (IMRT, VMAT) - Higher conformity, intentionally inhomogeneous dose in PTV.

    • High dose to smaller normal tissue volumes; larger normal tissue volumes receive low-dose bath.

Volume Effects & Clinical Tolerance

  • Volume irradiated can govern tolerance independently of intrinsic radiosensitivity.- Example: Skin – large-area exposure causes severe ulceration vs small-area exposure heals rapidly.

  • "Clinical tolerance" is ambiguously defined, varies by clinician & patient perceptions.

Power-Law Volume-Effect Models

  • Longstanding empirical rules:- Cube-root rule (Meyer 1939): Skin tolerance dose is proportional to A^-0.33 for field area A.

    • Jolles (1946): Tissue tolerance dose is proportional to V^-0.11 for volume V.

  • General power-law (iso-effect) model: Dv = D1 * v^-n

    • Dv: dose to fractional volume v = V/Vwhole giving same effect as whole-organ dose D_1.

    • Exponent n characterises volume sensitivity:

    • n < 0 for tumours (cold spots penalised).

    • n > 0 for normal tissue (hot spots penalised).

    • n = 0 → no volume effect (absolute maximum dose governs).

    • Small n → hot/cold spots not tolerated (e.g. spinal cord).

    Large n → organ tolerates inhomogeneity (e.g. liver).

  • Graphical data (Andrzej Niemierko, 2001) shows partial-volume tolerance for cord vs liver.

Functional Subunit (FSU) Concept (Withers 1988)

  • FSU = largest tissue unit recoverable from one surviving clonogenic cell.

  • Radiation can inactivate FSUs independently.

  • Organisation matters:- Parallel architecture

    • FSUs work independently; organ fails only if critical fraction lost.

    • Tolerance expressed via volume threshold (e.g. lung, kidney, liver).

    • Serial architecture

    • Failure of single FSU disables whole organ (e.g. spinal cord, optic nerve, intestine).

    • Most organs are hybrid (brain: parallel neurons but serial vasculature).

  • Explains paradox: Kidney (radiosensitive) can lose >50 % mass with little dysfunction; spinal cord (radio-resistant) fails with minimal damage.

Biological Plan-Evaluation Parameters (Modern TPS)

  • Planning systems now allow evaluation that merges dosimetric & radiobiological inputs:- Altered fractionation schedules.

    • Published normal-tissue tolerances (TD values).

    • Tissue/tumour radiosensitivity constants (alpha, beta).

    • Tumour doubling time / repopulation kinetics.

  • Algorithms output prospective probability metrics: Tumour Control Probability (TCP) & Normal Tissue Complication Probability (NTCP).

Probability Concepts in Radiotherapy

  • Success is inherently probabilistic, not deterministic.- Even after high total dose (e.g. 70 Gy) a very large tumour still harbours non-zero survival chance for a few clonogens.

  • Goal: Maximise TCP while minimising NTCP.

Tumour Control Probability (TCP)

  • Cell survival by Linear-Quadratic (LQ) model: S = e^-(alpha * D + beta * D^2)

    • alpha, beta: tissue-specific radiosensitivity constants.

    • D: total dose.

  • Assuming Poisson distribution of surviving clonogens: TCP = e^-(M * S)

    • M: initial number of clonogenic cells in tumour.

  • If beta approx 0 (single-strand killing dominant): S approx e^-(alpha * D) => TCP = e^-(M * e^-(alpha * D))

    • Produces typical sigmoid (S-shaped) dose–response curve.

  • Worked example from transcript:- Tumour contains 10^9 cells; alpha = 0.3 Gy^-1; beta approx 0.

    • Want TCP = 0.9 (90 %).

    • Set 0.9 = e^-(10^9 * e^-(0.3 * D)) => e^-(0.3 * D) = 10^-10.

    • Solve: D = ln(10^10) / 0.3 approx 69 Gy.

  • Expanded model (Nahum & Tait) for heterogeneous tumours & patient groups: Ns,tot = sum[i=1 to K] sum[j=1 to Y] N0,ij * e^-(alphai * Dj)

    • Accounts for spatial variation Dj and patient-specific alphai values.

Normal Tissue Complication Probability (NTCP)

  • Lyman–Kutcher–Burman (LKB) model: NTCP = (1 / sqrt(2 * pi)) * integral[-infinity to t] e^-(x^2 / 2) dx

    • t = (D - TD50(v)) / (m * TD50(v))

  • Volume scaling:

    • v = V / V_ref (fractional irradiated volume).

    • TD50(v) = TD50(1) * v^-n

  • Parameters:

    • TD_50(1) – uniform whole-organ dose giving 50 % complication rate.

    • n – volume-effect exponent (parallel/serial behaviour).

    • m – slope of NTCP curve (variability among patients).

  • Graph of TCP & NTCP vs dose illustrates competing probabilities; optimal therapeutic window often at region between curves.

Tissue Architecture Models Integrated with NTCP

Serial Model

  • Organ rendered non-functional if any one subunit exceeds tolerance.

  • NTCP approximated by maximum dose; reducing partial volume offers little benefit.

  • Clinical examples: spinal cord, optic nerve.

Parallel Model

  • Each sub-volume functions independently; organ fails when fraction damaged exceeds reserve.

  • NTCP increases with fraction of sub-volumes damaged (f_dam).

  • Clinical examples: lung, liver, kidney.

  • Planning strategy: “a little dose to a lot of tissue” vs “a lot dose to a little tissue”.

Summary & Clinical Decision-Making

  • Treatment Planning Systems (TPS) now output single-number metrics (TCP, NTCP) alongside DVHs.

  • Planners iterate beam geometry, modulation, and fractionation to maximise TCP and minimise NTCP, guided by:

    • Organ architecture (serial vs parallel).

    • Patient-specific biological inputs (alpha/beta, volume parameters).

    • Published tolerance data & clinical judgement.

  • Balancing act: Always a trade-off—higher tumour dose improves TCP but raises NTCP; modern biological models support more consistent, evidence-based decisions.